# Real Agent Cases

> A directory of 656 real, source-checked AI agent use cases. Each entry has the task, outcome, agent(s), evidence type, any claimed result (attributed to its source, not independently verified), a beginner-friendly flag, a reusable prompt (labeled original or recreated), and the original source link and date.

- [About](https://www.realagentcases.com/about/)
- [How we verify](https://www.realagentcases.com/how-we-verify/)
- [All agents](https://www.realagentcases.com/agent/)
- [Sitemap](https://www.realagentcases.com/sitemap.xml)
- [Full version with every prompt](https://www.realagentcases.com/llms-full.txt)

## Coding

### Import every saved Instagram food reel into a self-hosted recipe app with a pipeline Codex wrote
- URL: https://www.realagentcases.com/use-cases/import-every-saved-instagram-food-reel-into-a-self-hosted-recipe-app-with-a-pipe/
- Outcome: Karan Sharma gave Codex access to all their saved Instagram food reels and asked it to import them into their self-hosted Mealie recipe instance. Codex wrote a pipeline on the fly and ran each reel through it: yt-dlp pulls the audio, Whisper transcribes it, Gemini 3.8 Flash formats it as a recipe and it's sent to Mealie. Sharma calls it very useful.
- Agents: Codex
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/mrkaran_/status/2111523597349249308 (2026-10-17)
- Last verified: 2026-10-19

### Migrate two websites from WordPress to Wagtail with Claude Code
- URL: https://www.realagentcases.com/use-cases/migrate-two-websites-from-wordpress-to-wagtail-with-claude-code/
- Outcome: Over a weekend Claude Code migrated two sites from WordPress to Wagtail (PHP to Python), including server setup, data transfer, testing and fixes, and redesigned a homepage; the poster reports nothing broke and traffic and tools stayed in place.
- Agents: Claude Code
- Category: Coding
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/xamsor/status/2021513354431332677 (2026-02-11)
- Last verified: 2026-10-16

### Port a backend to Convex in four hours with Factory's Droid
- URL: https://www.realagentcases.com/use-cases/port-a-backend-to-convex-in-four-hours-with-factory-s-droid/
- Outcome: The poster reports that Factory's Droid ported their tRPC/Drizzle custom backend to Convex in about 4 hours, with some bugs left to fix.
- Agents: Factory Droid
- Category: Coding
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Dallenpyrah/status/1971996302746824729 (2025-09-27)
- Last verified: 2026-10-16

### Migrate a 50-screen app from Firebase to Supabase with Claude Code
- URL: https://www.realagentcases.com/use-cases/migrate-a-50-screen-app-from-firebase-to-supabase-with-claude-code/
- Outcome: Going feature by feature, the poster used Claude Code to generate database schemas, write queries and fix daily errors; they report migrating about 85% of a 45–50 screen Flutter app from Firebase to Supabase in 10–20 days.
- Agents: Claude Code
- Category: Coding
- Evidence: Described by poster
- Claimed result: ~85% migrated in 10–20 days (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/prayag_dalal/status/1963555583392714791 (2025-09-04)
- Last verified: 2026-10-16

### Build a blog feature by voice with Codex while cooking
- URL: https://www.realagentcases.com/use-cases/build-a-blog-feature-by-voice-with-codex-while-cooking/
- Outcome: Talking to Codex in voice mode while cooking dinner, the author got a Django newsletters archive with four importers and a pull request, then finished the rest with typed prompts.
- Agents: OpenAI Codex (ChatGPT desktop, voice mode)
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://simonwillison.net/2026/Oct/9/built-using-my-voice/ (2026-10-09)
- Last verified: 2026-10-16

### Migrate 180 microservices to a new framework version with Cursor agents
- URL: https://www.realagentcases.com/use-cases/migrate-180-microservices-to-a-new-framework-version-with-cursor-agents/
- Outcome: Brex reports that its self-correcting, self-hosted Cursor agents let one engineer migrate 180 microservices from Micronaut 3 to 4 in a month, which it calls a 32x per-capita speedup.
- Agents: Cursor (self-hosted agents)
- Category: Coding
- Evidence: Company case study
- Claimed result: 180 microservices migrated in one month (According to Brex; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.brex.com/journal/agent-automations-for-big-migrations (2026-07-22)
- Last verified: 2026-10-16

### Automate defect remediation across C/C++ codebases with Codex CLI
- URL: https://www.realagentcases.com/use-cases/automate-defect-remediation-across-c-c-codebases-with-codex-cli/
- Outcome: According to OpenAI's case study, Cisco's Codex CLI defect-repair workflow raised defect resolution throughput 10–15x, and Codex build optimization saved more than 1,500 engineering hours a month.
- Agents: OpenAI Codex CLI
- Category: Coding
- Evidence: Company case study
- Claimed result: 10–15x defect resolution throughput (According to OpenAI; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://openai.com/index/cisco/ (2026-05-27)
- Last verified: 2026-10-16

### Port an HTML5 parser from Python to JavaScript with Codex CLI
- URL: https://www.realagentcases.com/use-cases/port-an-html5-parser-from-python-to-javascript-with-codex-cli/
- Outcome: Codex CLI (GPT-5.2) ran for hours largely unattended and produced about 9,000 lines of dependency-free JavaScript passing 9,200 html5lib tests, in about 4.5 hours start to finish.
- Agents: OpenAI Codex CLI
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: Passes 9,200 html5lib tests (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Dec/15/porting-justhtml/ (2025-12-15)
- Last verified: 2026-10-16

### Upgrade a plugin to a new framework alpha with Codex CLI
- URL: https://www.realagentcases.com/use-cases/upgrade-a-plugin-to-a-new-framework-alpha-with-codex-cli/
- Outcome: A one-line `codex exec` loop read the Datasette 1.0a20 upgrade docs, fixed the failing tests, and got the datasette-checkbox plugin passing again.
- Agents: OpenAI Codex CLI
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Nov/6/upgrading-datasette-plugins/ (2025-11-06)
- Last verified: 2026-10-16

### Build a terminal-to-HTML sharing tool with Claude Code for web
- URL: https://www.realagentcases.com/use-cases/build-a-terminal-to-html-sharing-tool-with-claude-code-for-web/
- Outcome: Claude Code for web built a terminal-to-html tool that turns pasted terminal output into a shareable gist-preview link, recorded start to finish in an 11-minute video.
- Agents: Claude Code for web
- Category: Coding
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Oct/23/claude-code-for-web-video/ (2025-10-23)
- Last verified: 2026-10-16

### Get DeepSeek-OCR running on an NVIDIA Spark with Claude Code
- URL: https://www.realagentcases.com/use-cases/get-deepseek-ocr-running-on-an-nvidia-spark-with-claude-code/
- Outcome: With four prompts and about 40 minutes, Claude Code figured out the right ARM CUDA PyTorch wheels, got DeepSeek-OCR running on NVIDIA Spark hardware, and OCR'd a test image.
- Agents: Claude Code
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Oct/20/deepseek-ocr-claude-code/ (2025-10-20)
- Last verified: 2026-10-16

### Write and ship a complete Datasette plugin with Claude Code
- URL: https://www.realagentcases.com/use-cases/write-and-ship-a-complete-datasette-plugin-with-claude-code/
- Outcome: Claude Code (Sonnet 4.5) wrote every line of code and every test for the datasette-os-info plugin with minimal prompting, and the author shipped it.
- Agents: Claude Code
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Oct/8/claude-datasette-plugins/ (2025-10-08)
- Last verified: 2026-10-16

### Speed up feature delivery by moving a team to Cursor
- URL: https://www.realagentcases.com/use-cases/speed-up-feature-delivery-by-moving-a-team-to-cursor/
- Outcome: Security startup Jit moved its engineers from JetBrains to Cursor and, according to the article, delivered features 1.8–3x faster.
- Agents: Cursor
- Category: Coding
- Evidence: Article or interview
- Claimed result: 1.8x–3x productivity gains (Reported by LeadDev; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://leaddev.com/technical-direction/how-jit-overcame-developer-resistance-shift-cursor (2025-05-15)
- Last verified: 2026-10-16

### Generate a colophon page for a GitHub Pages site with Claude Code
- URL: https://www.realagentcases.com/use-cases/generate-a-colophon-page-for-a-github-pages-site-with-claude-code/
- Outcome: Claude Code wrote a script that builds a mobile-friendly colophon page listing every tool and its commit history, then wired it into the GitHub Pages deploy workflow.
- Agents: Claude Code
- Category: Coding
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Mar/11/using-llms-for-code/ (2025-03-11)
- Last verified: 2026-10-16

### Vibe-code a 3D town-building game with Claude Code
- URL: https://www.realagentcases.com/use-cases/vibe-code-a-3d-town-building-game-with-claude-code/
- Outcome: From a single informal prompt, Claude Code built a working 3D game for placing buildings and driving around in about four minutes; follow-up prompts added a fire truck, traffic, and fires.
- Agents: Claude Code
- Category: Coding
- Evidence: Article or interview
- Claimed result: Working game in about four minutes (Reported by One Useful Thing; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.oneusefulthing.org/p/speaking-things-into-existence (2025-03-11)
- Last verified: 2026-10-16

## Content

### Find repurposable clips in years of YouTube videos with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/find-repurposable-clips-in-years-of-youtube-videos-with-claude-cowork/
- Outcome: Given brand context, timestamped transcripts, example clips and performance data, Claude Cowork analyzed the content well and surfaced topics and hooks, saving hours of watching; but the clips it cut started mid-thought or lacked context or conclusions, even after several re-prompts.
- Agents: Claude Cowork
- Category: Content
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Great at analysis, but its own clip cuts were unusable (no context or conclusion) even after several prompts.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/ItsMinhaal/status/2092976763928539168 (2026-08-27)
- Last verified: 2026-10-16

### Write and publish a short podcast episode with Claude Code (needed edits)
- URL: https://www.realagentcases.com/use-cases/write-and-publish-a-short-podcast-episode-with-claude-code-needed-edits/
- Outcome: Claude Code wrote a short podcast script that the poster voiced with ElevenLabs and published via Anchor; the narration was clear but the script missed nuance and needed manual tweaks, so the poster rated it useful for drafts only.
- Agents: Claude Code
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The script missed nuance and needed manual edits; the poster called it useful for drafts only.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/AIorDIY/status/2052464770406854825 (2026-05-07)
- Last verified: 2026-10-16

### Make years of calls, podcasts and posts searchable with Claude Code
- URL: https://www.realagentcases.com/use-cases/make-years-of-calls-podcasts-and-posts-searchable-with-claude-code/
- Outcome: Claude Code organized years of calls, podcasts, tweets and presentations into a searchable database connected to AI, with a daily job that adds new content; asked for a masterclass outline, it drew on 224 chunks from 78 sources.
- Agents: Claude Code
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/andrehaykaljr/status/2021676546218434695 (2026-02-11)
- Last verified: 2026-10-16

### Script and edit a Google Vids video with ChatGPT Atlas
- URL: https://www.realagentcases.com/use-cases/script-and-edit-a-google-vids-video-with-chatgpt-atlas/
- Outcome: With Google Vids open, the poster asked ChatGPT Atlas in agent mode to create a video about the Atlas browser and says the scripting and editing were all done by Atlas.
- Agents: ChatGPT Atlas agent mode
- Category: Content
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/QRoonjha/status/1980740532990275755 (2025-10-21)
- Last verified: 2026-10-16

### Turn one video script into a newsletter, titles, thread and description with Claude Code subagents
- URL: https://www.realagentcases.com/use-cases/turn-one-video-script-into-a-newsletter-titles-thread-and-description-with-claud/
- Outcome: From one long-form video script, an orchestrator in Claude Code calls subagents that write a newsletter, generate titles, create thumbnails, write a thread and draft the YouTube description.
- Agents: Claude Code
- Category: Content
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/moritzkremb/status/1951636932846645260 (2025-08-02)
- Last verified: 2026-10-16

### Batch-enter design prompts into an image generator: Operator vs Project Mariner
- URL: https://www.realagentcases.com/use-cases/batch-enter-design-prompts-into-an-image-generator-operator-vs-project-mariner/
- Outcome: For an Etsy design shop, Operator entered a ChatGPT-made list of prompts into Ideogram to generate draft designs after the user logged in; Project Mariner couldn't get past Ideogram's captcha even with the user in control.
- Agents: OpenAI Operator, Project Mariner
- Category: Content
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Operator worked once the user logged in; Project Mariner could not get past the captcha.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/CodyThieling/status/1927128568045183167 (2025-05-26)
- Last verified: 2026-10-16

### Turn podcast highlights into timestamped clips with Claude Cowork + Claude in Chrome + Codex
- URL: https://www.realagentcases.com/use-cases/turn-podcast-highlights-into-timestamped-clips-with-claude-cowork-claude-in-chro/
- Outcome: Claude Cowork expanded the poster's top 10 podcast takeaways, then used Claude in Chrome to find the transcript and timestamps for each clip, and Codex downloaded them.
- Agents: Claude Cowork, Claude in Chrome, OpenAI Codex
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/daniel_mac8/status/2035782226571305039 (2026-03-22)
- Last verified: 2026-10-16

### Break writer's block on a book with Claude Code interviews
- URL: https://www.realagentcases.com/use-cases/break-writer-s-block-on-a-book-with-claude-code-interviews/
- Outcome: Claude Code reviewed a book draft, interviewed the author to fill gaps, and adapted his dictated answers into the manuscript in his own voice.
- Agents: Claude Code
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/marckohlbrugge/status/2022992628107870469 (2026-02-15)
- Last verified: 2026-10-16

### Debrief an 84-minute YouTube panel with Claude Code
- URL: https://www.realagentcases.com/use-cases/debrief-an-84-minute-youtube-panel-with-claude-code/
- Outcome: Claude Code installed yt-dlp, pulled and cleaned the captions of an 84-minute Gurley/Gerstner/Nadella talk, wrote focused debriefs, then packaged it as a reusable skill.
- Agents: Claude Code
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/vha14/status/2014848771503620499 (2026-01-23)
- Last verified: 2026-10-16

### Generate weekly AI content ideas with an n8n agent
- URL: https://www.realagentcases.com/use-cases/generate-weekly-ai-content-ideas-with-an-n8n-agent/
- Outcome: An n8n agent researches top AI news via Brave Search, generates content from templates, and emails a weekly idea pack.
- Agents: n8n agent
- Category: Content
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/_juliettech/status/1983231404914929686 (2025-10-28)
- Last verified: 2026-10-16

### Turn stroller-walk voice notes into a published article with Claude Code
- URL: https://www.realagentcases.com/use-cases/turn-stroller-walk-voice-notes-into-a-published-article-with-claude-code/
- Outcome: A non-developer fed rambling voice notes to Claude Code, which organized them into research themes, wrote a full article in her voice and made LinkedIn versions.
- Agents: Claude Code
- Category: Content
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.lennysnewsletter.com/p/everyone-should-be-using-claude-code (2025-10-14)
- Last verified: 2026-10-16

### Prep podcast research packets with an n8n agent
- URL: https://www.realagentcases.com/use-cases/prep-podcast-research-packets-with-an-n8n-agent/
- Outcome: Emily Lambert built an app on an n8n agent that researches a podcast topic, pulls news and tweets, then emails a script/outline to the team.
- Agents: n8n agent
- Category: Content
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/emilylambert/status/1932083209380274313 (2025-06-09)
- Last verified: 2026-10-16

### Turn a blog post into a Quarto slide deck with Claude Code
- URL: https://www.realagentcases.com/use-cases/turn-a-blog-post-into-a-quarto-slide-deck-with-claude-code/
- Outcome: Claude Code read one of the author's blog posts and turned it into a Quarto slide deck in three prompts.
- Agents: Claude Code
- Category: Content
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kyle_e_walker/status/1931850344855691498 (2025-06-08)
- Last verified: 2026-10-16

## Data

### Graph monthly traffic deaths for two states with Grok
- URL: https://www.realagentcases.com/use-cases/graph-monthly-traffic-deaths-for-two-states-with-grok/
- Outcome: The poster says he asked Grok to graph Iowa traffic deaths per month for the last five years, then California, to check whether a roadside death counter looked low; he shared the charts.
- Agents: Grok
- Category: Data
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/TuelTaylor/status/2103482706315321689 (2026-09-25)
- Last verified: 2026-10-16

### Build a support data pipeline and dashboard in under an hour with Devin
- URL: https://www.realagentcases.com/use-cases/build-a-support-data-pipeline-and-dashboard-in-under-an-hour-with-devin/
- Outcome: At Eight Sleep, a support manager who used to need the data team for dashboards built an end-to-end pipeline and dashboard herself with Devin in under an hour, according to Devin's customer story.
- Agents: Devin
- Category: Data
- Evidence: Company case study
- Claimed result: Under an hour (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://devin.ai/customers/eight-sleep (2026-07-16)
- Last verified: 2026-10-16

### Set up a Google Analytics dashboard with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/set-up-a-google-analytics-dashboard-with-claude-in-chrome/
- Outcome: Claude in Chrome spent 45 minutes building a Google Analytics dashboard, then couldn't work out how to drag one element, so it undid all of its work and told the poster it couldn't do it.
- Agents: Claude in Chrome
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. After 45 minutes it got stuck on a drag step and undid everything instead of saving progress.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/GonzoHacker/status/2038363718640156786 (2026-03-29)
- Last verified: 2026-10-16

### Turn a year of bank statements into a business expense spreadsheet with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/turn-a-year-of-bank-statements-into-a-business-expense-spreadsheet-with-claude-c/
- Outcome: Claude Cowork extracted and categorized business expenses from a year of statements across several accounts into a well-organized spreadsheet; after some back-and-forth the poster fixed a handful of items by hand.
- Agents: Claude Cowork
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/MikeMoves/status/2036708713562513630 (2026-03-25)
- Last verified: 2026-10-16

### Build a credit risk scorecard in Excel with Claude in Excel
- URL: https://www.realagentcases.com/use-cases/build-a-credit-risk-scorecard-in-excel-with-claude-in-excel/
- Outcome: Given a credit risk dataset and a brief for an Excel-native scorecard, Claude in Excel built a polished 28-sheet workbook but skipped the variable clustering, ran the model in JavaScript and reported 0.706 AUC on training data with no split; ChatGPT followed the method and got 0.627 on test data.
- Agents: Claude (Excel)
- Category: Data
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Looked impressive but broke the brief.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/akshen121/status/2031029965605699699 (2026-03-09)
- Last verified: 2026-10-16

### Find smarter buying patterns in office supply order history with Claude Code
- URL: https://www.realagentcases.com/use-cases/find-smarter-buying-patterns-in-office-supply-order-history-with-claude-code/
- Outcome: Alex Hillman had Claude Code go through Indy Hall's paper products order history to suggest buying patterns that avoid costly last-minute orders.
- Agents: Claude Code
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/alexhillman/status/2020352927848923387 (2026-02-08)
- Last verified: 2026-10-16

### Analyze seven years of credit card statements with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/analyze-seven-years-of-credit-card-statements-with-claude-cowork/
- Outcome: From seven years of monthly credit card statement PDFs in zip files, Claude Cowork returned an Excel file with 5,726 extracted, categorized transactions plus analysis; issues were vendor transaction IDs and some miscategorized vendors it wasn't told how to handle.
- Agents: Claude Cowork
- Category: Data
- Evidence: Screenshots or logs
- Claimed result: 5,726 transactions categorized (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/_simonsmith/status/2018380086311800948 (2026-02-02)
- Last verified: 2026-10-16

### Sort a Google Sheet in a custom priority order with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/sort-a-google-sheet-in-a-custom-priority-order-with-claude-in-chrome/
- Outcome: Asked to sort a Google Sheet in a custom order (critical, high, medium, low), Claude for Chrome kept trying to sort it alphabetically instead, per the poster's screenshot.
- Agents: Claude in Chrome
- Category: Data
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Repeatedly sorted alphabetically instead of in the custom order.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/AaronBergman18/status/1983740965660844115 (2025-10-30)
- Last verified: 2026-10-16

### Build a cohort retention table in Excel with Copilot Agent Mode
- URL: https://www.realagentcases.com/use-cases/build-a-cohort-retention-table-in-excel-with-copilot-agent-mode/
- Outcome: Given a long table of customers by period, Excel's Copilot Agent Mode built a color-coded cohort retention table that DataCamp's reviewer found accurate to the data; it also added a useless chart he had to delete.
- Agents: Microsoft Copilot Agent Mode (Excel)
- Category: Data
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.datacamp.com/blog/agent-mode-in-excel (2025-10-05)
- Last verified: 2026-10-16

### Fill missing industry details in a customer sheet with Zapier Agents
- URL: https://www.realagentcases.com/use-cases/fill-missing-industry-details-in-a-customer-sheet-with-zapier-agents/
- Outcome: UK energy company egg set up a Zapier Agent that finds customers with missing industry details in a HubSpot-fed Google Sheet, searches for the information and updates the sheet; Zapier says it now takes about five minutes and egg no longer pays a contractor for it.
- Agents: Zapier Agents
- Category: Data
- Evidence: Company case study
- Claimed result: About 5 minutes; contractor no longer needed (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://zapier.com/blog/egg-sparks-clean-energy-sales-with-ai/ (2024-10-07)
- Last verified: 2026-10-16

### Let a whole team self-serve data via an agent-written schema guide
- URL: https://www.realagentcases.com/use-cases/let-a-whole-team-self-serve-data-via-an-agent-written-schema-guide/
- Outcome: Claude Code wrote a markdown guide to the company's data model and business terms from SQLMesh context and past queries; Hex's agent now handles the team's data requests. The team had posted a data analyst role, and the poster now thinks the role is 'completely dead'.
- Agents: Claude Code, Hex agent
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/dylangans/status/2047768695678578723 (2026-04-24)
- Last verified: 2026-10-16

### Merge provider price CSVs into one comparison table with Claude Code
- URL: https://www.realagentcases.com/use-cases/merge-provider-price-csvs-into-one-comparison-table-with-claude-code/
- Outcome: Claude Code read several local CSVs of model prices from different providers and produced one organized comparison table grouped by provider.
- Agents: Claude Code
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/S0laceGao/status/2036850598855766030 (2026-03-25)
- Last verified: 2026-10-16

### Automate a daily analytics summary to Slack with Claude Code
- URL: https://www.realagentcases.com/use-cases/automate-a-daily-analytics-summary-to-slack-with-claude-code/
- Outcome: Claude Code set up a GitHub Actions cron that pulls PostHog data at 8am, summarizes it in Python and posts to Slack; the first report arrived the next morning.
- Agents: Claude Code
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/PranavDate/status/2031210925806530779 (2026-03-10)
- Last verified: 2026-10-16

### Turn X bookmarks and likes into a searchable knowledge base
- URL: https://www.realagentcases.com/use-cases/turn-x-bookmarks-and-likes-into-a-searchable-knowledge-base/
- Outcome: The poster scraped all their X bookmarks and likes with Claude in Chrome and other tools, cleaned the linked pages with Firecrawl, and built a RAG chatbot over everything they had saved.
- Agents: Claude in Chrome, Vercel Agent Browser
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/sheeki03/status/2013898161782481048 (2026-01-21)
- Last verified: 2026-10-16

### Analyze every entry of a huge JSONL file with Claude Code subagents
- URL: https://www.realagentcases.com/use-cases/analyze-every-entry-of-a-huge-jsonl-file-with-claude-code-subagents/
- Outcome: Claude Code previewed the data, trimmed fields, defined a schema and ran hundreds of Haiku subagents over ~45 minutes so every entry in a large JSONL file was analyzed.
- Agents: Claude Code
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/sean_moriarity/status/2012220803023073539 (2026-01-16)
- Last verified: 2026-10-16

### Compare Claude Code, Codex and Gemini CLI on a 97k-message chat dataset
- URL: https://www.realagentcases.com/use-cases/compare-claude-code-codex-and-gemini-cli-on-a-97k-message-chat-dataset/
- Outcome: The poster gave the same 97k-message WeChat dataset to Claude Code, Codex and Gemini CLI as analysts and published results with an evaluation table.
- Agents: Claude Code, OpenAI Codex CLI, Gemini CLI
- Category: Data
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/tylerxiety/status/1988989909429022850 (2025-11-13)
- Last verified: 2026-10-16

### Reproduce an economics paper's findings from replication data with Claude
- URL: https://www.realagentcases.com/use-cases/reproduce-an-economics-paper-s-findings-from-replication-data-with-claude/
- Outcome: Given only the paper and its replication archive, Claude Sonnet 4.5 converted the STATA code to Python and reproduced the findings; the author spot-checked the results and GPT-5 Pro confirmed them.
- Agents: Claude (Sonnet 4.5)
- Category: Data
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.oneusefulthing.org/p/real-ai-agents-and-real-work (2025-09-29)
- Last verified: 2026-10-16

## Marketing

### Write a drama-style ad script modeled on a winning ad with Manus
- URL: https://www.realagentcases.com/use-cases/write-a-drama-style-ad-script-modeled-on-a-winning-ad-with-manus/
- Outcome: Manus wrote a drama-format script for one of the brand's top historical winning ad concepts, modeled on a winning drama ad from a beauty brand; the poster says it hit 5 ROAS right away.
- Agents: Manus
- Category: Marketing
- Evidence: Described by poster
- Claimed result: 5 ROAS off the bat (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/thankyouecom/status/2107259289224053025 (2026-10-05)
- Last verified: 2026-10-16

### Find car brands without a web app and email them a pitch with Manus
- URL: https://www.realagentcases.com/use-cases/find-car-brands-without-a-web-app-and-email-them-a-pitch-with-manus/
- Outcome: Manus searched for car brands without a web app, wrote an email and sent it to them; some replied and the poster built one an app.
- Agents: Manus
- Category: Marketing
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/JerryJenasom/status/2102306956203442534 (2026-09-22)
- Last verified: 2026-10-16

### Manage Google Ads with Claude Code connected to analytics and the website
- URL: https://www.realagentcases.com/use-cases/manage-google-ads-with-claude-code-connected-to-analytics-and-the-website/
- Outcome: With Google Ads, Google Analytics and the Webflow site connected to a Claude Code project, the poster adds negative keywords quickly, tracks ad performance, and found that mobile wasn't converting because the page showed nothing relevant above the fold.
- Agents: Claude Code
- Category: Marketing
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/michaelsxacc/status/2087519918803566662 (2026-08-12)
- Last verified: 2026-10-16

### Qualify scraped leads against your ideal customer before cold email with Claude Code
- URL: https://www.realagentcases.com/use-cases/qualify-scraped-leads-against-your-ideal-customer-before-cold-email-with-claude/
- Outcome: Claude Code reads each scraped company's website and checks it against the poster's ideal customer profile, dropping about 30% of leads before outreach; the poster says reply rates roughly doubled and they booked 100+ meetings in 3 months.
- Agents: Claude Code
- Category: Marketing
- Evidence: Described by poster
- Claimed result: 100+ meetings booked in 3 months (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ZaidAD_/status/2062271127213236350 (2026-06-03)
- Last verified: 2026-10-16

### Make a PDF brochure from my website with Manus
- URL: https://www.realagentcases.com/use-cases/make-a-pdf-brochure-from-my-website-with-manus/
- Outcome: Asked to create a professional PDF brochure from the poster's website, Manus browsed and analyzed the site but delivered what was essentially copy-paste into a plain Google Doc; after feedback it only added a coloured first-page background and a different font, and the poster cancelled.
- Agents: Manus
- Category: Marketing
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Delivered plain copy-pasted text; a retry changed only the background and font.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: original
- Source: https://x.com/jfonseca81/status/2022683037402538446 (2026-02-14)
- Last verified: 2026-10-16

### Write a competitor content-marketing report with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/write-a-competitor-content-marketing-report-with-chatgpt-agent/
- Outcome: In Lisa Peyton's test, ChatGPT Agent Mode took 18 minutes to produce a 12-page narrative competitive-intelligence report on a list of competitors, with activity summaries, trend analysis, 31 cited sources and recommendations for content strategy.
- Agents: ChatGPT agent
- Category: Marketing
- Evidence: Article or interview
- Claimed result: 12-page report in 18 minutes; 31 sources (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head AI agent review (ChatGPT agent vs Manus), not a standalone test.
- Source: https://lisapeyton.com/the-ai-agent-showdown-chatgpt-vs-manus-ai-for-content-marketing-competitive-intelligence/ (2025-09-23)
- Last verified: 2026-10-16

### Download a gated marketing report and summarize it with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/download-a-gated-marketing-report-and-summarize-it-with-chatgpt-agent/
- Outcome: SE Ranking's reviewer asked ChatGPT agent to get HubSpot's 2025 State of Marketing report; it found the page, filled in the gated form, downloaded the report and wrote a summary without help, but Google blocked it from signing in to email the summary because it was using an unsupported browser.
- Agents: ChatGPT agent
- Category: Marketing
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. It downloaded and summarized the report on its own, but couldn't email the summary: Google blocked the sign-in from the agent's browser.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://seranking.com/blog/chatgpt-agent-digital-marketing-review/ (2025-07-24)
- Last verified: 2026-10-16

### Write a cited industry report on AI in B2B visual marketing with Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/write-a-cited-industry-report-on-ai-in-b2b-visual-marketing-with-gemini-deep-research/
- Outcome: Lisa Peyton gave Gemini, ChatGPT and Perplexity Deep Research the same prompt for a report on generative AI in B2B visual marketing; Gemini's roughly 9,500-word report was rated highly accurate with 95 distinct citations, though she found it more theoretical than practical, with fewer pricing details and implementation examples than ChatGPT's.
- Agents: Gemini Deep Research
- Category: Marketing
- Evidence: Article or interview
- Claimed result: About 9,500 words; 95 citations (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://lisapeyton.com/deep-deeper-deepest-which-ai-is-the-best-deep-research-tool-for-marketers/ (2025-03-22)
- Last verified: 2026-10-16

### Research competitors for a board update with ChatGPT Deep Research
- URL: https://www.realagentcases.com/use-cases/research-competitors-for-a-board-update-with-chatgpt-deep-research/
- Outcome: Section's chief of staff gave ChatGPT Deep Research and Gemini Deep Research the same competitive-analysis prompt for its AI coaching product, written from a board member's point of view; she rated ChatGPT's output 4.5/5, saying it surfaced less obvious competitors such as Skillsoft and DataCamp and gave exactly what she asked for, while Gemini's report scored 2.5/5 for being too high level.
- Agents: ChatGPT Deep Research
- Category: Marketing
- Evidence: Article or interview
- Claimed result: Output rated 4.5/5 (Gemini 2.5/5) (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.sectionai.com/blog/chatgpt-vs-gemini-deep-research (2025-03-14)
- Last verified: 2026-10-16

### Submit a product to ~300 directories and launch sites with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/submit-a-product-to-300-directories-and-launch-sites-with-claude-in-chrome/
- Outcome: In one day, Claude in Chrome filled directory and launch-platform forms 5 tabs at a time; 134 of about 300 submissions went through, with the human handling logins and captchas.
- Agents: Claude in Chrome
- Category: Marketing
- Evidence: Described by poster
- Claimed result: 134 submissions went through (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/_ksl_dev/status/2105293870766444553 (2026-09-30)
- Last verified: 2026-10-16

### Fix a Product Hunt X handle across launches and comments with a browser agent
- URL: https://www.realagentcases.com/use-cases/fix-a-product-hunt-x-handle-across-launches-and-comments-with-a-browser-agent/
- Outcome: A browser agent corrected the poster's X handle from johnrushx to johnrush across all Product Hunt launches and hundreds of comments.
- Agents: Browser agent
- Category: Marketing
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/johnrush/status/2087198876688412929 (2026-08-11)
- Last verified: 2026-10-16

### Audit 90 days of Meta ads from a CSV export with Claude Code
- URL: https://www.realagentcases.com/use-cases/audit-90-days-of-meta-ads-from-a-csv-export-with-claude-code/
- Outcome: Claude Code flagged 14 money-losing creatives, found that 7 of the top 10 ads opened with a question, and spotted 3 ad sets cannibalizing each other.
- Agents: Claude Code
- Category: Marketing
- Evidence: Described by poster
- Claimed result: 14 money-losing creatives flagged (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/propialis/status/2041264181106692318 (2026-04-06)
- Last verified: 2026-10-16

### Automate a daily cold-email pipeline with Claude Code
- URL: https://www.realagentcases.com/use-cases/automate-a-daily-cold-email-pipeline-with-claude-code/
- Outcome: A weekend-built Claude Code pipeline scrapes leads, scores them against the ICP, writes spintax copy and launches the Instantly campaign with one command; the poster says it saves about 2 hours each morning.
- Agents: Claude Code
- Category: Marketing
- Evidence: Described by poster
- Claimed result: 2 hours back every morning (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Jason_scales1/status/2039976196339859727 (2026-04-03)
- Last verified: 2026-10-16

### Build a B2B SaaS press list of 40–50 journalists with Manus
- URL: https://www.realagentcases.com/use-cases/build-a-b2b-saas-press-list-of-40-50-journalists-with-manus/
- Outcome: Manus produced a Notion-ready press list of journalists covering mid-market SaaS/enterprise software with contacts, recent articles and pitch angles; the author says work that used to take ~6 hours took about 25 minutes, with some contact verification still needed.
- Agents: Manus
- Category: Marketing
- Evidence: Article or interview
- Claimed result: 6 hours of work done in 25 minutes (Reported by The AI Enterprise; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.theaienterprise.io/p/ai-agents-research (2025-10-21)
- Last verified: 2026-10-16

### Generate 60+ AI-consulting sales prospects with Manus
- URL: https://www.realagentcases.com/use-cases/generate-60-ai-consulting-sales-prospects-with-manus/
- Outcome: A Manus prompt for mid-market AI-implementation prospects returned about 73 qualified companies with decision-maker contacts and outreach notes in a Notion/spreadsheet export.
- Agents: Manus
- Category: Marketing
- Evidence: Article or interview
- Claimed result: 73 prospects found (Reported by The AI Enterprise; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.theaienterprise.io/p/ai-agents-research (2025-10-21)
- Last verified: 2026-10-16

### Find Southeast AI events for meetup promotion with Manus
- URL: https://www.realagentcases.com/use-cases/find-southeast-ai-events-for-meetup-promotion-with-manus/
- Outcome: Manus found 34 upcoming AI/tech events in North Carolina and the Southeast suitable for promoting a meetup, exported to Notion with contacts and opportunity type, cutting an afternoon of work to about 20 minutes.
- Agents: Manus
- Category: Marketing
- Evidence: Article or interview
- Claimed result: 34 relevant events found (Reported by The AI Enterprise; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.theaienterprise.io/p/ai-agents-research (2025-10-21)
- Last verified: 2026-10-16

### Weekly X analytics review and growth coaching with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/weekly-x-analytics-review-and-growth-coaching-with-chatgpt-agent/
- Outcome: Every Monday ChatGPT agent pulls the poster's X analytics, studies which posts worked, compares profile visits with follower growth, and suggests bio and pinned-post changes.
- Agents: ChatGPT agent
- Category: Marketing
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AIwithJai/status/1967708746752327852 (2025-09-15)
- Last verified: 2026-10-16

### Fix Google Ads conversion tracking with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/fix-google-ads-conversion-tracking-with-chatgpt-agent/
- Outcome: A solo creator had ChatGPT agent log into her Google Ads account and fix a conversion-tracking issue.
- Agents: ChatGPT agent
- Category: Marketing
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/evielync/status/1953929013703193013 (2025-08-08)
- Last verified: 2026-10-16

### Build an SEO content calendar with Manus
- URL: https://www.realagentcases.com/use-cases/build-an-seo-content-calendar-with-manus/
- Outcome: Manus studied the author's products and writing, researched SEO and keywords, and produced a content calendar in a 30+ minute run that the author says cost about $10.
- Agents: Manus
- Category: Marketing
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.alara.blog/ai-agents-experiences-with-manus-operator-and-more/ (2025-05-31)
- Last verified: 2026-10-16

### Generate qualified sales leads with Zapier Agents
- URL: https://www.realagentcases.com/use-cases/generate-qualified-sales-leads-with-zapier-agents/
- Outcome: According to Zapier's customer story, Slate Studios produced 2,000+ qualified leads in one month with a single Zapier Agent that also handled data reformatting and competitive research.
- Agents: Zapier Agents
- Category: Marketing
- Evidence: Company case study
- Claimed result: 2,000+ leads generated (According to Zapier; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://zapier.com/blog/how-slate-generates-2000-leads-with-zapier-agents/ (2025-04-30)
- Last verified: 2026-10-16

### Automate PR lead discovery from news with Zapier Agents
- URL: https://www.realagentcases.com/use-cases/automate-pr-lead-discovery-from-news-with-zapier-agents/
- Outcome: According to Zapier's customer story, NisonCo grew weekly leads from 270 to nearly 400 (+48%) and saved about $30,000 a year with an agent that scans news for new businesses in its target industries.
- Agents: Zapier Agents
- Category: Marketing
- Evidence: Company case study
- Claimed result: 48% increase in leads (According to Zapier; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://zapier.com/blog/how-nisonco-fuels-business-growth-with-zapier-agents/ (2025-02-18)
- Last verified: 2026-10-16

## Ops

### Have Claude Cowork prepare late-invoice reminder drafts every Monday night, with a report of who owes what
- URL: https://www.realagentcases.com/use-cases/have-claude-cowork-prepare-late-invoice-reminder-drafts-every-monday-night-with/
- Outcome: Michel, in San Diego, works in a tight-margin business with basic back-office software, and sending late-invoice reminders was repetitive manual work. Claude Cowork walked them through setting up an agent that logs in every Monday night, checks overdue accounts receivable and matches customer names against a Google Sheet of contacts. Using Michel's email template, it adds the contact name, invoice number, amount and due date, puts the emails in the Drafts folder and sends Michel a report of which customers were included, what each owes and the totals. Michel reviews the drafts and sends them, saving about an hour every Monday night.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://app.therundown.ai/community/posts/81f4a019-1b41-4b86-868c-c287b3064249 (2026-10-13)
- Last verified: 2026-10-19

### Redline a vendor's standard OEM agreement with Claude for the first three rounds, then have a lawyer do the final check
- URL: https://www.realagentcases.com/use-cases/redline-a-vendor-s-standard-oem-agreement-with-claude-for-the-first-three-rounds/
- Outcome: jedberg regularly negotiates OEM agreements with vendors that mostly use the same template and draw the same objections. They tried Claude first, and their lawyer was impressed that it found things the lawyer had missed. jedberg then used Claude to redline that type of contract, and the other side remarked that the lawyer had been surprisingly thorough. The lawyer did the final check and by then basically said it looked good, and jedberg saved a lot of money by doing the first three rounds with Claude.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=50060191 (2026-10-12)
- Last verified: 2026-10-19

### Sort a year of mixed Amazon business purchases into QuickBooks categories with Claude, including the fiddly split transactions
- URL: https://www.realagentcases.com/use-cases/sort-a-year-of-mixed-amazon-business-purchases-into-quickbooks-categories-with-c/
- Outcome: badlibrarian runs a business and sometimes buys things on Amazon. At tax time they used to search for the big business items and categorise them in QuickBooks Online until they gave up, using a janky browser plugin and two monitors full of Excel windows. Now Claude does it and doesn't give up, despite QuickBooks being slow. It will even click 19 times to split a transaction because they'd put a router and shampoo in the same Amazon cart one night in February 2025.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=50047430 (2026-10-11)
- Last verified: 2026-10-19

### Have Claude draft replies to a small brand's customer messages, warranty claims included, with a human approving each one
- URL: https://www.realagentcases.com/use-cases/have-claude-draft-replies-to-a-small-brand-s-customer-messages-warranty-claims-i/
- Outcome: idopmstuff runs several small consumer brands. For the most recent one they acquired, customers got in touch through Facebook and Instagram comments and messages as well as email, so they had Claude build an app that pulls those messages in as they arrive and drafts a reply, including any actions needed, for them to edit or send. Their edits feed back into its playbook, and it has got pretty good. It handles warranty claims (ask for a photo of the damage and the order number, create a replacement order, email a confirmation) and most common questions. Every reply is still approved by a human, though they plan to let it handle common cases on its own soon.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49732493 (2026-09-16)
- Last verified: 2026-10-19

### Get a 0.75% to 12.75% unemployment insurance rate jump fixed retroactively with ChatGPT filing the missing wage report
- URL: https://www.realagentcases.com/use-cases/get-a-0-75-to-12-75-unemployment-insurance-rate-jump-fixed-retroactively-with-ch/
- Outcome: mattmaroon runs a seasonal business and noticed their state unemployment insurance rate had jumped from 0.75% to 12.75%. Instead of hours on the state website and on hold, they asked ChatGPT what had happened. It worked out that they were being charged a penalty rate because their new payroll provider doesn't file wage reports for the one quarter with no wages, and without a report the business counted as delinquent. ChatGPT filed the correct wage report and went to the support site to request a retroactive rewrite, which was granted, saving many thousands of dollars without hitting the 5-hour usage limit.
- Agents: ChatGPT
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49467734 (2026-08-27)
- Last verified: 2026-10-19

### Do a small business's quarterly BAS books from bank, Stripe, PayPal and Wise statements with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/do-a-small-business-s-quarterly-bas-books-from-bank-stripe-paypal-and-wise-state/
- Outcome: Sridhar Katakam used Claude Cowork to go through the statements and transactions from several bank accounts, Stripe, PayPal and Wise and prepare business expense and income spreadsheets. Their accountant reviewed them and said they were good enough that Katakam no longer needed the accountant's services for BAS and could file it themselves, a saving of about $530 AUD per quarter. The accountant also suggested having Claude do the same for the whole year, and said the fee for filing the income tax return would come down.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/srikat/status/2086643189776253312 (2026-08-10)
- Last verified: 2026-10-19

### Ask Claude whether to convert an LLC to an S Corp, and miss that NYC taxes make it cost more
- URL: https://www.realagentcases.com/use-cases/ask-claude-whether-to-convert-an-llc-to-an-s-corp-and-miss-that-nyc-taxes-make-i/
- Outcome: mikeocool asked Claude about converting their LLC to an S Corp. Claude said it was a great idea and estimated how much they would save in taxes. Their accountant said it would actually cost more, because of how New York City taxes S Corps. They hadn't told Claude they lived in NYC, because it didn't occur to them that it mattered. Their takeaway: tax questions are full of details like this, often subtler than where you live.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Failure. Claude said the switch would save tax; the accountant said NYC rules meant it would cost more.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49140304 (2026-08-02)
- Last verified: 2026-10-19

### Recreate and resend an urgent client invoice from a phone in a roller coaster queue with Claude
- URL: https://www.realagentcases.com/use-cases/recreate-and-resend-an-urgent-client-invoice-from-a-phone-in-a-roller-coaster-qu/
- Outcome: A client called Alper Ortac on a family holiday at Europa Park needing an invoice urgently for their end-of-year accounts. With no laptop, Ortac opened Claude on their phone in a roller coaster queue, attached the last invoice and asked it to recreate it. Claude wrote and ran a Python script that produced a replica of the PDF, then changed the invoice number, billing period and costs as asked. Ortac drafted the email, attached the new invoice and sent it in roughly 20 minutes, right before boarding.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/alperortac/status/2074849882624389428 (2026-07-08)
- Last verified: 2026-10-19

### Turn a PTA's bank exports into year-end charity accounts in four minutes with Claude
- URL: https://www.realagentcases.com/use-cases/turn-a-pta-s-bank-exports-into-year-end-charity-accounts-in-four-minutes-with-cl/
- Outcome: Pete Hickerton, chair of their school's PTA, wanted to save the treasurer an evening on the year-end accounts. The job meant exporting bank transactions, categorising everything, reconciling and formatting it for the Charity Commission. They gave the bank exports to Claude, and four minutes later had a set of accounts ready to upload. Claude explained what it had done, flagged where it had checked the formulas and showed the reconciliation. Hickerton and the treasurer went through it looking for errors and found none.
- Agents: Claude
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/petehickerton_our-pta-treasurer-was-about-to-lose-an-evening-activity-7470829883548549120-OTsS (2026-06-11)
- Last verified: 2026-10-19

### Turn a box of 63 crumpled receipts into a QuickBooks batch-upload sheet with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/turn-a-box-of-63-crumpled-receipts-into-a-quickbooks-batch-upload-sheet-with-cla/
- Outcome: John Fong had a box of receipts that had sat on their desk for 1.5 years, some crumpled, some with faded ink, and some restaurant meals split across two receipts with the tip on its own. They started a Claude Cowork project called AP Agent, photographed 63 receipts into a folder and asked for an Excel sheet they could batch upload to QuickBooks. They took a shower and 8 minutes later came back to a completed file with every receipt allocated to an account and amount. It flagged 3 receipts it had only partly read, asking them to check those, and 3 that weren't valid expenses under CRA (Canada Revenue Agency) rules.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/johnfong_this-is-the-reality-of-many-small-businesses-activity-7470552025060274176-hQzu (2026-06-10)
- Last verified: 2026-10-19

### Run a restaurant's back office through an OpenClaw agent: supplier price-change alerts, invoice uploads and a fairer tip split
- URL: https://www.realagentcases.com/use-cases/run-a-restaurant-s-back-office-through-an-openclaw-agent-supplier-price-change-a/
- Outcome: Christopher Thépot runs restaurant operations in Madrid and spent a month using OpenClaw. The agent answers point-of-sale questions over WhatsApp and combines supplier invoices, sales data, recipes and inventory counts to track waste and what's missing. It flags supplier price spikes; two pricing mistakes caught so far directly affected the P&L. It also pulls supplier invoices from their inbox into the accounting software's scanner, saving one to two hours a week, though they're still fine-tuning that. They used it to design and test a new tip policy in under an hour: 80% stays with the shift and 20% goes to a pool split by hours worked.
- Agents: OpenClaw
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/christopher-th%C3%A9pot-03217391_openclaw-activity-7452697032458334209-S_6n (2026-04-22)
- Last verified: 2026-10-19

### Build a five-question Google Form with an AI browser agent: Gemini handed it back, Claude finished it (partial)
- URL: https://www.realagentcases.com/use-cases/build-a-five-question-google-form-with-an-ai-browser-agent-gemini-handed-it-back/
- Outcome: Irena Palamani Xhurxhi asked Gemini to build a Google Form. It got partway, then stopped and asked Xhurxhi to take over because it couldn't change the question type in the interface. Xhurxhi gave the same task to Claude in the browser and went back to other work; when they returned, the form was done, with all five questions, correct types, descriptions and required fields. Asked about it, Claude said it had struggled with the same parts, trying clicks, JavaScript workarounds and keyboard navigation, but kept going rather than handing it back.
- Agents: Gemini, Claude in Chrome
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The first agent gave up partway; a second agent finished the form.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ipxhurxhi_i-asked-gemini-to-build-a-google-form-for-activity-7441919095320768512-r88r (2026-03-23)
- Last verified: 2026-10-19

### Turn a photo of a cafe's daily inventory sheet into shopping lists for three suppliers with a custom GPT
- URL: https://www.realagentcases.com/use-cases/turn-a-photo-of-a-cafe-s-daily-inventory-sheet-into-shopping-lists-for-three-sup/
- Outcome: Cole Morgan runs a small cafe with no single major supplier, so they drive to three stores. Staff fill in two paper inventory forms each day and upload them to Slack, and turning those into a checklist for each store took 2-10 minutes. An off-the-shelf OCR tool that worked well cost over $250 a month, and a Python script they had ChatGPT write wasn't accurate enough. So they built a custom GPT, 'C&K Inventory Manager', trained on a spreadsheet of every item and its vendor. Now they upload the sheet photo and get a checkbox list grouped by vendor, ready to paste into Obsidian. It asks about items missing from the sheet, such as whole milk. Accuracy improved with GPT-4o with no change to the instructions, though it sometimes asks too many questions and they have to tell it to just output the list.
- Agents: ChatGPT
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.colemorgan.com/posts/how-i-automated-my-cafe-inventory-ordering-with-chatgpt/ (2025-01-12)
- Last verified: 2026-10-19

### Stop a runaway cloud bill and win a billing credit with ChatGPT's help
- URL: https://www.realagentcases.com/use-cases/stop-a-runaway-cloud-bill-and-win-a-billing-credit-with-chatgpt-s-help/
- Outcome: A poster says ChatGPT helped them stop a runaway Google Cloud bill, prepare the evidence and, under their supervision, talk to a real Google Cloud support agent. They report a TWD 4,043 credit, with the remaining TWD 883.50 still under appeal.
- Agents: ChatGPT
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The poster reports part of the bill credited and the rest still under appeal.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Izzo55688/status/2085399964478582822 (2026-08-06)
- Last verified: 2026-10-17

### Find a list of missing invoices across work and personal email with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-a-list-of-missing-invoices-across-work-and-personal-email-with-chatgpt-agen/
- Outcome: Testing ChatGPT agent, a poster gave it a list of invoices he owed his operations assistant, and he says it found all of them in his personal and work email. He also notes it booked accommodation successfully (they had to contact the hotel to cancel the test booking) and estimates about a 40% end-to-end completion rate on complex tasks for now.
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/dawidnaude_weve-been-testing-chatgpt-agent-since-its-activity-7358371337981026305-Nk4h (2025-08-05)
- Last verified: 2026-10-17

### Build a Power Automate Desktop flow with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/build-a-power-automate-desktop-flow-with-claude-cowork/
- Outcome: After Copilot suggested a subflow that didn't exist in Power Automate Desktop, the poster asked Claude Cowork, which took over and built the flow on their machine.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/magicj3/status/2105499283126407656 (2026-10-01)
- Last verified: 2026-10-16

### Submit a hosting provider bug report from log warnings with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/submit-a-hosting-provider-bug-report-from-log-warnings-with-claude-in-chrome/
- Outcome: Using the poster's logged-in browser, Claude in Chrome submitted a bug report to DigitalOcean after finding warnings in the logs.
- Agents: Claude in Chrome
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ndbroadbent/status/2104499860368138397 (2026-09-28)
- Last verified: 2026-10-16

### Split, rename and sort about 200 scanned invoices with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/split-rename-and-sort-about-200-scanned-invoices-with-claude-cowork/
- Outcome: After the poster scanned almost 200 invoices, Claude Cowork split them out, renamed them, sorted them into folders and created a summary Excel file.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/IlyaPion/status/2102783493482709261 (2026-09-23)
- Last verified: 2026-10-16

### Find, negotiate and review a small office lease with Codex
- URL: https://www.realagentcases.com/use-cases/find-negotiate-and-review-a-small-office-lease-with-codex/
- Outcome: Codex found a 74 sq ft commercial lease near the poster's kids' school, negotiated the terms and reviewed the contract; the poster found it cheaper than a coworking desk.
- Agents: Codex
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kolencherry/status/2094983316306858030 (2026-09-02)
- Last verified: 2026-10-16

### Prepare an Indian income tax return (ITR-3) with Claude Code
- URL: https://www.realagentcases.com/use-cases/prepare-an-indian-income-tax-return-itr-3-with-claude-code/
- Outcome: Given tax documents and a year of bank statements, Claude Code reconciled the figures against the poster's AIS (annual information statement), caught 2 mistakes that could have triggered notices, and computed the tax; the poster then filed.
- Agents: Claude Code
- Category: Ops
- Evidence: Described by poster
- Claimed result: 2 mistakes caught (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/m_jamal_jm/status/2075517698319286312 (2026-07-10)
- Last verified: 2026-10-16

### QA-test a department chatbot and file Jira bugs with Claude Cowork and Chrome
- URL: https://www.realagentcases.com/use-cases/qa-test-a-department-chatbot-and-file-jira-bugs-with-claude-cowork-and-chrome/
- Outcome: Using Claude Cowork with the Chrome extension, the poster had it navigate a new chatbot like a user and produce a report of what worked, what needed improvement, bugs, repro steps and fixes; after the poster reproduced the bugs, Claude created Jira tickets assigned to the right teammates.
- Agents: Claude Cowork, Claude in Chrome
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Kole_Sam/status/2062964099974369751 (2026-06-05)
- Last verified: 2026-10-16

### Reconcile crypto trades across chains and post them to Xero with Claude Code
- URL: https://www.realagentcases.com/use-cases/reconcile-crypto-trades-across-chains-and-post-them-to-xero-with-claude-code/
- Outcome: Claude Code tracked on-chain swaps across 10+ assets, produced a monthly gain/loss breakdown with reconciled exchange rates, then built a Xero connection and posted journal entries.
- Agents: Claude Code
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ksaitor/status/2042163099600638171 (2026-04-09)
- Last verified: 2026-10-16

### Match a tax-prep checklist against a receipts folder with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/match-a-tax-prep-checklist-against-a-receipts-folder-with-claude-cowork/
- Outcome: The poster says matching a spreadsheet of items for their accountant against a folder of receipts usually took 4–6 hours, and that with Claude Cowork it took about an hour.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: 4–6 hour job done in about an hour (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/bharat/status/2034068158123020569 (2026-03-18)
- Last verified: 2026-10-16

### Prepare and file overdue German VAT returns with Claude Cowork and Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/prepare-and-file-overdue-german-vat-returns-with-claude-cowork-and-claude-in-chr/
- Outcome: From a folder of invoice and receipt PDFs, Claude Cowork prepared quarterly accounts; after the poster logged in to the Elster tax portal, Claude in Chrome found, filled and submitted the Q3 2025 VAT forms.
- Agents: Claude Cowork, Claude in Chrome
- Category: Ops
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/EastlondonDev/status/2020826097890242582 (2026-02-09)
- Last verified: 2026-10-16

### Turn six months of receipts into a categorized spreadsheet with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/turn-six-months-of-receipts-into-a-categorized-spreadsheet-with-claude-cowork/
- Outcome: Pointed at a folder, Claude Cowork organized 6 months of receipts into a categorized spreadsheet with monthly breakdowns.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ChuckYouSuck/status/2014502869336989810 (2026-01-23)
- Last verified: 2026-10-16

### Organize a messy SEO work folder and rename screenshots with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/organize-a-messy-seo-work-folder-and-rename-screenshots-with-claude-cowork/
- Outcome: Claude Cowork asked how the poster wanted 27 unstructured files organized, created analytics-data, blog-posts and screenshots folders, looked at each screenshot and renamed it by content (e.g. feature-text-to-sql-converter.jpg); the poster says it took about 30 seconds.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Shown on video
- Claimed result: 27 files sorted in about 30 seconds (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/mustafaergisi/status/2011432413390344579 (2026-01-14)
- Last verified: 2026-10-16

### Write outgoing wire invoices with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/write-outgoing-wire-invoices-with-chatgpt-agent/
- Outcome: The poster has ChatGPT agent write all their outgoing wire invoices and says it saves 5 minutes three times a week.
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Described by poster
- Claimed result: 5 minutes, 3 times a week (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ChrisGPT/status/2007210390141506043 (2026-01-02)
- Last verified: 2026-10-16

### Do year-end bookkeeping and Xero reconciliation with Claude Code
- URL: https://www.realagentcases.com/use-cases/do-year-end-bookkeeping-and-xero-reconciliation-with-claude-code/
- Outcome: Claude Code categorized a folder of receipts, cross-referenced exported bank CSVs against Xero to find duplicates and missing transactions, used the browser extension to report on reconciliation items, and produced a cost-cutting report; the poster says this replaced roughly $2–3K of accounting-firm work.
- Agents: Claude Code, Claude in Chrome
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: ~2–3K accounting-firm fee avoided (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/amirmxt/status/2004658010212499713 (2025-12-26)
- Last verified: 2026-10-16

### Withdraw about 1,000 old LinkedIn invitations with Perplexity's agent
- URL: https://www.realagentcases.com/use-cases/withdraw-about-1-000-old-linkedin-invitations-with-perplexity-s-agent/
- Outcome: Perplexity Pro's agent mode kept getting tripped up by LinkedIn's infinite scroll, and when asked to withdraw invitations it was painfully slow and got stuck after about 5; the poster had Perplexity write a Cursor prompt for a Selenium script instead.
- Agents: Perplexity agent mode
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Stuck after about 5 withdrawals; the poster switched to a script.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/BenyaminHolley/status/1971210280395808873 (2025-09-25)
- Last verified: 2026-10-16

### Complete a 31-step web form: ChatGPT agent vs Perplexity Comet
- URL: https://www.realagentcases.com/use-cases/complete-a-31-step-web-form-chatgpt-agent-vs-perplexity-comet/
- Outcome: On a 31-step form with buttons, dropdowns and fields, Comet failed after 3–5 steps, while ChatGPT agent completed all 31 and corrected its own mistakes by checking against the steps.
- Agents: ChatGPT agent, Perplexity Comet
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Comet failed after 3–5 steps; ChatGPT agent finished all 31.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/rohitdotmittal/status/1951450652867174754 (2025-08-02)
- Last verified: 2026-10-16

### Summarize a 10-page lease renewal before signing with Grok
- URL: https://www.realagentcases.com/use-cases/summarize-a-10-page-lease-renewal-before-signing-with-grok/
- Outcome: An X user says that to renew an office lease with the landlord, they asked Grok to summarize the 10+ pages of legal terms instead of going over them with an attorney, and it took less than 30 minutes to fully review the details themselves. The post includes a screenshot of Grok's summary of the shopping-center lease amendment, covering the term extension and rent schedule.
- Agents: Grok
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/affinity892/status/1875295564302303235 (2025-01-03)
- Last verified: 2026-10-16

### Enter retail-arbitrage receipts into a spreadsheet with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/enter-retail-arbitrage-receipts-into-a-spreadsheet-with-claude-cowork/
- Outcome: Claude Cowork weekly processes a folder of RA receipts into a spreadsheet with buy cost, store and item; the poster still double-checks but says admin time is way down.
- Agents: Claude Cowork
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ZachAltmyer/status/2033904740359147632 (2026-03-17)
- Last verified: 2026-10-16

### Draft government-solicitation proposals in 25 minutes with n8n AI workflows
- URL: https://www.realagentcases.com/use-cases/draft-government-solicitation-proposals-in-25-minutes-with-n8n-ai-workflows/
- Outcome: According to n8n's case study, Field Aerospace's self-hosted n8n AI workflows bring a proposal draft that used to take about two weeks of work from three or four people to an 80% starting point in roughly 25 minutes, and replaced two legacy tools, eliminating approximately $30,000 in annual software costs (the case study's headline gives $22,000 a year).
- Agents: n8n (AI-powered workflows)
- Category: Ops
- Evidence: Company case study
- Claimed result: Proposal time cut from 2 weeks to 25 minutes (According to n8n; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://n8n.io/case-studies/field-aerospace/ (2026-03-03)
- Last verified: 2026-10-16

### Build a weekly business KPI dashboard workflow with Claude Code + n8n
- URL: https://www.realagentcases.com/use-cases/build-a-weekly-business-kpi-dashboard-workflow-with-claude-code-n8n/
- Outcome: Claude Code built a complex n8n workflow for a client's weekly KPI dashboard in about 15 minutes; the workflow then runs unattended on n8n.
- Agents: Claude Code, n8n
- Category: Ops
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://landauai.com/ai-agents-claude-code-vs-workflow-tools-n8n-which-should-automate-your-processes/ (2026-02-27)
- Last verified: 2026-10-16

### Auto-coach sales reps from call recordings with Zapier Agents
- URL: https://www.realagentcases.com/use-cases/auto-coach-sales-reps-from-call-recordings-with-zapier-agents/
- Outcome: According to Zapier's customer story, Healthie's call-summary and coaching agents save each of about 20 reps and CSMs 2–3 hours a week, more than 60 hours a week across the team.
- Agents: Zapier Agents
- Category: Ops
- Evidence: Company case study
- Claimed result: 60+ hours a week saved (According to Zapier; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://zapier.com/blog/healthie-saves-60-hours-per-week-with-ai-agents/ (2025-11-25)
- Last verified: 2026-10-16

### Sort and rename messy invoices for taxes with Claude Code
- URL: https://www.realagentcases.com/use-cases/sort-and-rename-messy-invoices-for-taxes-with-claude-code/
- Outcome: Claude Code reads each file in a messy invoice folder, renames it to 'YYYY-MM-DD Vendor - Invoice - ProductOrService.pdf' and files it in the right folder.
- Agents: Claude Code
- Category: Ops
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.lennysnewsletter.com/p/everyone-should-be-using-claude-code (2025-10-14)
- Last verified: 2026-10-16

### Categorize 419 ledger transactions into an Excel template with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/categorize-419-ledger-transactions-into-an-excel-template-with-chatgpt-agent/
- Outcome: Given an Excel ledger template and 419 transactions, ChatGPT agent asked for help on ambiguous merchants, then keyed in all ~400 rows and took 4 minutes.
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: Done in 4 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/jasononfirms/status/1963993526515192040 (2025-09-05)
- Last verified: 2026-10-16

### Allocate a 30-employee payroll journal across departments with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/allocate-a-30-employee-payroll-journal-across-departments-with-chatgpt-agent/
- Outcome: ChatGPT agent mapped a payroll CSV into a 30-employee allocation spreadsheet across 6 departments, making 180 entries and summing employer-paid columns in under 4 minutes.
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: 180 entries in under 4 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/jasononfirms/status/1963628292873109898 (2025-09-04)
- Last verified: 2026-10-16

### Fill an accelerator application with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/fill-an-accelerator-application-with-chatgpt-agent/
- Outcome: ChatGPT agent filled an accelerator application using the poster's context and chat memory; the poster rated it 8/10, and also had it research vendors into a spreadsheet (7/10).
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/mahtabsoin/status/1949203910922764689 (2025-07-26)
- Last verified: 2026-10-16

### Diagnose a Kubernetes pod-scheduling outage with Claude Code
- URL: https://www.realagentcases.com/use-cases/diagnose-a-kubernetes-pod-scheduling-outage-with-claude-code/
- Outcome: Anthropic's Data Infrastructure team gave Claude Code dashboard screenshots, and it traced the problem to pod IP exhaustion and supplied the commands to add a new IP pool; Anthropic says this saved about 20 minutes during an outage.
- Agents: Claude Code
- Category: Ops
- Evidence: Company case study
- Claimed result: 20 minutes saved during an outage (According to Anthropic; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://claude.com/resources/articles/how-anthropic-teams-use-claude-code (2025-07-24)
- Last verified: 2026-10-16

### Three-way match purchase orders against delivery and invoice docs with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/three-way-match-purchase-orders-against-delivery-and-invoice-docs-with-chatgpt-a/
- Outcome: In about 4 minutes ChatGPT agent checked that all 5 POs had matching delivery and invoice documents with no variances, combined the PDFs, and wrote a short report.
- Agents: ChatGPT agent
- Category: Ops
- Evidence: Screenshots or logs
- Claimed result: Done in around 4 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/jand__/status/1946283011445784991 (2025-07-18)
- Last verified: 2026-10-16

## Personal

### Catch an error in a 4th-grader's textbook answer key, which ChatGPT first got wrong too (failure)
- URL: https://www.realagentcases.com/use-cases/catch-an-error-in-a-4th-grader-s-textbook-answer-key-which-chatgpt-first-got-wro/
- Outcome: Carlos Mendoza's 4th-grade son marked only B and F as correct in a Santillana exercise on numbers up to the billions. The platform's solution marked B, D and F, but D said '9,615 rounded to the billions is 10,000', which is false. When Mendoza asked ChatGPT to review it, it first said Santillana was right, reading 'billions' as if it were 'thousands'. Only when Mendoza pointed to the exact word did ChatGPT acknowledge the error. The son had been right from the beginning.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. ChatGPT agreed with the wrong answer key until the parent pointed to the exact word.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/carlosatal/status/2111517945386156108 (2026-10-17)
- Last verified: 2026-10-19

### Get a recycled phone number's spam for a previous owner sorted with Meta's Muse, which bot-blocked support kept stopping (failure)
- URL: https://www.realagentcases.com/use-cases/get-a-recycled-phone-number-s-spam-for-a-previous-owner-sorted-with-meta-s-muse/
- Outcome: Amanda Smith pays $25 a month for a Visible plan and gets a lot of spam calls and texts asking for someone called Andrew, probably the number's previous owner. Visible has no phone line, so Smith asked Muse to check they were formally listed as the number's owner and to find where the number was linked to Andrew. Within seconds it found the number on a data broker page. But Visible blocks bots, and its guest chat wouldn't send Muse's message, then blocked it; taking over the browser led to a 403 error. Muse drafted a message for Smith to paste into the Visible app, and advised requesting removal from WhitePages and FastBackgroundCheck, though the WhitePages form wouldn't let Smith click Next. Visible's advanced support was still pending. Smith found Muse acted more like a chatbot giving instructions, though testing it made Smith finally deal with the task.
- Agents: Muse
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Bot blocking stopped Muse at the carrier and the data broker form; the problem wasn't fixed at the time of writing.
- Prompt type: Original (as published in the source)
- Source: https://www.cnet.com/tech/services-and-software/meta-muse-contact-phone-provider-ai-agent-blocked/ (2026-10-16)
- Last verified: 2026-10-19

### Find a cheaper Medicare drug plan with ChatGPT after a premium rise, then confirm it in Medicare's Plan Finder
- URL: https://www.realagentcases.com/use-cases/find-a-cheaper-medicare-drug-plan-with-chatgpt-after-a-premium-rise-then-confirm/
- Outcome: A Topretirements blogger's Medicare Part D drug plan wrote to say the 2027 premium would rise from $98.30 to $126 a month. They asked ChatGPT for cheaper options, giving their Florida ZIP code, their medications and preferred pharmacies. It suggested several, including a plan at $5.30 a month in their area, $1,448.40 a year less in premiums. They didn't switch on ChatGPT's word alone: as their insurance agent also advised, they checked it in Medicare's official Plan Finder, which confirmed the switch made sense, and enrolled. They get most prescriptions through the VA, so the cheapest premium suited them; they note others need to compare total yearly cost.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.topretirements.com/blog/i-saved-1448-on-my-medicare-drug-plan-you-might-save-big-too/ (2026-10-16)
- Last verified: 2026-10-19

### Work out why a meeting was missed and set up a tested email-to-calendar check with ChatGPT and Codex
- URL: https://www.realagentcases.com/use-cases/work-out-why-a-meeting-was-missed-and-set-up-a-tested-email-to-calendar-check-wi/
- Outcome: Dave W Plummer missed an important appearance and asked the ChatGPT desktop app, which has access to their Gmail, to work out how it happened. It found Plummer had noon in mind, the time of an earlier prep call, while the event started at 7:50 a.m. Pacific, with emails also mentioning 8:50 a.m. Mountain. When Codex then reconstructed the email trail, it found the organiser had cancelled the calendar invite sent to Plummer's regular address while sorting out a Teams login mismatch, and the later confirmation email had no new invite. With Codex, Plummer then set up a recurring agent that reads email threads, including changes and cancellations, adds missing calendar entries with the source email and link, and flags contradictions. Test alerts at first didn't notify on the iPhone; after checking alarms, sync and Focus settings, a test finally rang. It now checks every ten minutes while the Mac is awake; its first historical review was still running.
- Agents: ChatGPT, Codex
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/davepl1968/status/2110796715179876531 (2026-10-15)
- Last verified: 2026-10-19

### Order a one-off meal prep delivery with Grok Bot and end up on a recurring subscription by mistake (failure)
- URL: https://www.realagentcases.com/use-cases/order-a-one-off-meal-prep-delivery-with-grok-bot-and-end-up-on-a-recurring-subsc/
- Outcome: ApoStructura asked Grok Bot to place a one-off meal prep order. Instead, it signed them up for a recurring meal prep subscription, costing £57 ($75) extra. They posted to claim the Grok Bot 'we'll make you whole' guarantee.
- Agents: Grok Bot
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Grok Bot signed the user up for a recurring subscription instead of a one-off order.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/ApoStructura/status/2108270660275609612 (2026-10-08)
- Last verified: 2026-10-19

### Track 186 applications, find missing keywords and review interview notes with Claude to land an offer
- URL: https://www.realagentcases.com/use-cases/track-186-applications-find-missing-keywords-and-review-interview-notes-with-cla/
- Outcome: Margareth Felicia Gono, an international student who needed a visa, started job hunting scattershot. Gono built an application tracker with Claude, had it pull the main skills and keywords from each job description and compare them with the résumé, and used it as a career coach, which led to three résumés for analytics, sales and marketing, and consulting. Where interviewers allowed recording, Gono ran sentiment analysis on the notes to see where Gono hesitated or used filler words. The tracker showed 186 applications, 44 cold outreaches, 19 interviews, six take-home assignments, three final rounds and one offer, as a senior associate at DoorDash.
- Agents: Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/job-seeker-used-claude-track-search-improve-interviews-2026-10 (2026-10-04)
- Last verified: 2026-10-19

### Get a $199 membership refund for slow test results by citing the provider's own agreement with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-199-membership-refund-for-slow-test-results-by-citing-the-provider-s-own-a/
- Outcome: Gokul N.'s blood test results from Superpower took 20 days. Gokul asked ChatGPT to check the refund policy, look at the dates and email the concierge. The email cited Section 14(a) of the membership agreement, set out the timeline with Superpower's own support replies, asked them to confirm when results had actually been available, and added a fallback asking for an exception if a different policy applied. Two days later Superpower confirmed a full $199 membership refund.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/gokul5544_chatgpt-helped-me-get-a-199-refund-my-prompt-activity-7511198770441900032-3PQt (2026-09-30)
- Last verified: 2026-10-19

### Push the vet to run a test for FIP after Claude kept flagging it from a sick rescue cat's results
- URL: https://www.realagentcases.com/use-cases/push-the-vet-to-run-a-test-for-fip-after-claude-kept-flagging-it-from-a-sick-res/
- Outcome: thobes and their spouse took in a scared stray cat found in Elysian Park on July 4th. Weeks later the cat stopped eating and declined fast; after-hours ER care in LA cost $7,500 for two days of life support before any diagnosis, and treatment moved to a clinic in Tijuana. thobes ran each batch of test results through Claude, which kept saying to look into FIP and to push the vet to test the stomach fluid for it, though the first vet waved it off. When the Tijuana clinic was asked for the extra FIP test they first refused ('AI is not a vet'), then agreed. It came back positive, the cat got a prescription and is expected to live. thobes says not to trust medical decisions to AI entirely.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Thobey_Campion/status/2103674822022898055 (2026-09-26)
- Last verified: 2026-10-19

### Find a year of forgotten FSA-eligible purchases across three retailers and file 15 claims with an AI agent in Cursor
- URL: https://www.realagentcases.com/use-cases/find-a-year-of-forgotten-fsa-eligible-purchases-across-three-retailers-and-file/
- Outcome: Ramesh Raman dislikes filing reimbursements, so they pulled 12 months of receipts from three retailer accounts and had an LLM classify every item as FSA-eligible or not. They checked the ones it got wrong themselves until every eligible item had a PDF receipt, then pointed Cursor at their reimbursement portal to file the claims. From 600 transactions it submitted 15 claims worth $242.87, pausing for approval before every upload; they reviewed everything at the end. The first upload took about 2 minutes and the last under 30 seconds. It took about an hour, for a chore they'd put off for months, and mostly recovered small forgotten buys like Tylenol.
- Agents: Cursor
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/rameshdotr_i-dislike-filing-reimbursements-so-i-decided-activity-7507951871009906688-L_P6 (2026-09-22)
- Last verified: 2026-10-19

### Find out a spouse's Tax ID could be got for free in a minute instead of paying an agent 10k with ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-out-a-wife-s-tax-id-could-be-got-for-free-in-a-minute-instead-of-paying-an/
- Outcome: Mr Jasper's spouse needed a Tax ID, and the agent they contacted quoted 10k and asked for the spouse's NIN to process it the next day. Mr Jasper asked ChatGPT about it and how long it takes, and found it was something they could do themselves in a few minutes. They opened the official portal with the spouse's NIN and the Tax ID was ready in less than one minute, saving the 10k fee.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Mr_Jay888/status/2102092870295449710 (2026-09-21)
- Last verified: 2026-10-19

### Plan an international family move in three weeks with a Claude project, and have it build the Notion tracker and calendar reminders
- URL: https://www.realagentcases.com/use-cases/plan-an-international-family-move-in-three-weeks-with-a-claude-project-and-have/
- Outcome: Iliyana Stareva had three weeks to move a family of four from Spain back to the Netherlands. They opened a Claude project called "Move Back to the Netherlands", gave it their history, details about their kids and what was already sorted (they still had Dutch bank accounts and a phone number), and asked for a full plan covering housing, schools and childcare, moving both cars, taxes and municipality paperwork. In five minutes the plan surfaced a few things they had forgotten. Claude then built a tracker in Notion with owner, deadline and category for each task plus a "this week" view, and put key dates into Google Calendar, including a call to the municipality. Some items didn't apply and some deadlines moved once checked against their work calendar, so they spent ten more minutes adjusting it.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://iliyanastareva.substack.com/p/we-moved-back-to-the-netherlands (2026-09-15)
- Last verified: 2026-10-19

### Check how much is left on a car loan with ChatGPT Work and Grok Bot, and give up at the lender's login (failure)
- URL: https://www.realagentcases.com/use-cases/check-how-much-is-left-on-a-car-loan-with-chatgpt-work-and-grok-bot-and-give-up/
- Outcome: John Helmuth asked both ChatGPT Work and Grok Bot how much time was left on their Ford Explorer loan with Ford Credit. Both first searched email for a statement, found none and went to the Ford Credit site, then got stuck at the login. ChatGPT's secure form didn't work and its cloud browser never showed a keyboard; Grok Bot handed over its cloud browser with a working keyboard, but after Helmuth entered their email and password it stayed stuck on the sign-in screen. After 5 minutes on both, Helmuth logged in on their phone's browser with a passkey and saw the loan terms in about 30 seconds.
- Agents: ChatGPT Work, Grok Bot
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Neither agent got past the lender's sign-in; the poster checked on their phone in 30 seconds.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/johnhelmuth_/status/2096596752790954437 (2026-09-06)
- Last verified: 2026-10-19

### Sort 204 unread emails into action labels with ChatGPT, approving every change first
- URL: https://www.realagentcases.com/use-cases/sort-204-unread-emails-into-action-labels-with-chatgpt-approving-every-change-fi/
- Outcome: Ary Aranguiz connected ChatGPT to Gmail and asked it to create four labels: ACTION/Reply, ACTION/Decision, ACTION/Review and STATUS/Waiting. ChatGPT then reviewed 204 unread emails from the past seven days and proposed which belonged under each label, separating people waiting for a reply, invitations needing a decision, information to review and routine mail needing no action. It noticed the original label set was incomplete and recommended adding ACTION/Review. Aranguiz approved the proposals, and only then did it create the label and apply the labels. It didn't send, delete, archive, forward or mark anything as read.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/aryaranguiz_i-just-gave-chatgpt-access-to-my-gmail-activity-7501654785218129920-xvrY (2026-09-04)
- Last verified: 2026-10-19

### Get a pet insurer's $359 'earned premium' collections threat settled at $118 by escalating to the state regulator with Claude
- URL: https://www.realagentcases.com/use-cases/get-a-pet-insurer-s-359-earned-premium-collections-threat-settled-at-118-by-esca/
- Outcome: After two pet insurance policies lapsed, Dimitri Sudomoin was billed $359 in 'earned premium' and spent two months being threatened with collections. Weeks of emails with the insurer's customer team got nowhere and a supervisor closed the thread saying they had gone as far as they could. With Claude planning the campaign, Sudomoin filed a complaint with the insurer's home-state regulator in New York and emailed two of its executives. The insurer folded in three and a half hours: the bill was settled at $118 as a 'special exception', with a written promise that nothing ever went to a credit bureau.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.sudomoin.com/p/consumer-claims (2026-09-04)
- Last verified: 2026-10-19

### Get a refused warranty exchange offered after Claude found the company's own arbitration clause (partial)
- URL: https://www.realagentcases.com/use-cases/get-a-refused-warranty-exchange-offered-after-claude-found-the-company-s-own-arb/
- Outcome: Dimitri Sudomoin's $1,166 robot mop failed under warranty. After eleven weeks of 'forwarding this to our technical team', the company refused to honour the warranty and offered a final 46% partial refund. Claude's first idea, a federal warranty law angle, looked weak, but reading the terms of service it found the company's own arbitration clause: under the rules of the forum the company picked, Sudomoin's filing cost was capped at $250 and the company would bear the rest of the fees. Three days after Sudomoin filed, the company offered the exchange it had previously refused. Sudomoin describes this one as still in play.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. An exchange was offered after the filing, but the post says this dispute is still in play.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.sudomoin.com/p/consumer-claims (2026-09-04)
- Last verified: 2026-10-19

### Run your side of a burst-pipe home insurance claim with Claude: receipts, logs, policy checks and pushing back on delays (partial)
- URL: https://www.realagentcases.com/use-cases/run-your-side-of-a-burst-pipe-home-insurance-claim-with-claude-receipts-logs-pol/
- Outcome: Two months after a water pipe burst in their house, wellpast is using Claude to run their side of the claim with a large insurer. It reads the policy, organises and files scans of receipts for additional living expenses (ALE), reads and flags messages from the agents, logs all correspondence, work estimates and invoices, points out when the insurer is delaying or not keeping promises, and challenges coverage decisions based on the policy. The adjuster gets very tidy, documented receipts and logs. The delays have been bad enough that Claude now insists they should file a complaint with the state insurance department. Without Claude, they say, they'd probably have given up from fatigue.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The claim was still in progress; Claude recommended a regulator complaint over delays.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49536267 (2026-09-02)
- Last verified: 2026-10-19

### Catch a passport-validity problem five days before a Europe trip by asking Claude what else to do pre-trip
- URL: https://www.realagentcases.com/use-cases/catch-a-passport-validity-problem-five-days-before-a-europe-trip-by-asking-claud/
- Outcome: Sarah Silbert was using Claude to choose checked luggage for a trip to Germany and asked, almost as an afterthought, what else needed doing before the trip. Among the obvious items, Claude said to check the passport had three months of validity past the return date. It didn't: it expired about two months after the return flight, one month short, five days before departure. Inside the 14-day window for an urgent passport agency appointment, Silbert got a slot the next day, paid the $60 expedite fee and picked up a new passport the following day, using the State Department website as the official source.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://pointspath.kit.com/posts/the-least-clever-ai-prompt-was-the-one-that-saved-my-trip (2026-08-29)
- Last verified: 2026-10-19

### Trace a check-engine code and a tank that won't fill to a stuck valve with Grok, and fix it for $120
- URL: https://www.realagentcases.com/use-cases/trace-a-check-engine-code-and-a-tank-that-won-t-fill-to-a-stuck-valve-with-grok/
- Outcome: Christopher Wipper's mechanic wanted $169 just to troubleshoot a check engine light. AutoZone's free OBD2 scanner gave code P0458 (purge control valve circuit low), and the car wouldn't take fuel: it backed up in the fill tube. With no YouTube videos for the problem, Wipper used Grok to troubleshoot step by step: the air drain case on top of the fill port was clear, but the canister close valve's solenoid was stuck closed. Wipper had the replacement part in hand in 20 minutes and fixed it for $120 in less than an hour.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/SGTWipper1Each/status/2093820992091005386 (2026-08-29)
- Last verified: 2026-10-19

### Catch a returned $2,350 insurance reimbursement check buried in 80k unread emails with Grok Bot
- URL: https://www.realagentcases.com/use-cases/catch-a-returned-2-350-insurance-reimbursement-check-buried-in-80k-unread-emails/
- Outcome: Within five minutes of setting up Grok Bot, a poster with about 80k unread emails had it flag one message from four days earlier: the insurance company said their windshield replacement reimbursement check had been returned by the post office and needed a new address. The poster says it was the only email the bot flagged and that it saved them $2350.
- Agents: Grok Bot
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/thecyberfam/status/2091549225918042236 (2026-08-23)
- Last verified: 2026-10-19

### Get the kindergarten's daily menu every morning by having Claude Code find the school site's hidden data feed
- URL: https://www.realagentcases.com/use-cases/get-the-kindergarten-s-daily-menu-every-morning-by-having-claude-code-find-the-s/
- Outcome: Nikunj Kothari's daughter had just started kindergarten, and the daily meal was posted on a website with messy, unstructured data. Kothari asked Claude Code to find the site's API by looking at its network requests; it turned out to be unauthenticated, and Claude Code worked out the right format and handed it to the family's existing Hermes bot. Now their home bot tells them each morning what's for breakfast and lunch so they can pack food to match.
- Agents: Claude Code
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/nikunj/status/2090884422178627624 (2026-08-21)
- Last verified: 2026-10-19

### Rebuild a pool's plumbing for about $1k instead of a $6k quote, with ChatGPT drawing a labelled parts diagram
- URL: https://www.realagentcases.com/use-cases/rebuild-a-pool-s-plumbing-for-about-1k-instead-of-a-6k-quote-with-chatgpt-drawin/
- Outcome: joenot443 was quoted $6k to fix their pool's plumbing and rebuilt it themselves over a few days with YouTube and GPT, for about $1k of materials. It was their first outdoor plumbing job and they'd never touched PVC. They say they wouldn't have had the confidence from YouTube alone: specific diagnostic help and a full diagram made the difference, and they had it prepare an SVG of the whole assembly with every measurement and part labelled.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49256359 (2026-08-11)
- Last verified: 2026-10-19

### Compare five homeowner's insurance quotes with ChatGPT and Claude, and learn to check insurers' financial ratings, not BBB
- URL: https://www.realagentcases.com/use-cases/compare-five-homeowner-s-insurance-quotes-with-chatgpt-and-claude-and-learn-to-c/
- Outcome: Natasha Hsi got five homeowner's insurance quotes for a complicated house. They differed on exclusions, wind and named-storm deductibles, and admitted versus surplus-lines insurers, and Hsi found them hard to follow. They fed all five plus their current renewal into ChatGPT and Claude and ran scenarios such as a hurricane loss and a total loss. When Hsi suggested starting with BBB ratings, both pushed back: BBB tracks complaints, while financial strength ratings show whether an insurer can pay after a storm. Hsi took the questions back to their original agent, who returned a better-structured quote at a more reasonable cost with none of the weakly rated insurers.
- Agents: ChatGPT, Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/natashahsi_i-got-5-homeowners-insurance-quotes-today-activity-7493018407139713024-lu9W (2026-08-11)
- Last verified: 2026-10-19

### Check a $3,802 surgery bill against insurance with Gemini and get a $269 consultation charge removed
- URL: https://www.realagentcases.com/use-cases/check-a-3-802-surgery-bill-against-insurance-with-gemini-and-get-a-269-consultat/
- Outcome: Financial planner Mike Palmer's itemised statement for a surgery was $3,802 in confusing codes and offsetting entries; insurance covered $1,390, leaving $2,412 out of pocket. Palmer saved the bill details as a PDF without personal information and asked Gemini what the codes meant and whether the coverage was correct. Gemini explained each line, confirmed the write-offs showed in-network pricing, and flagged an office consultation fee as likely contested, since insurers usually bundle a same-day evaluation into the procedure. There had been no separate consultation, so Palmer called the provider to dispute it, and they removed the $269 fee.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.kiplinger.com/personal-finance/guide-to-using-ai-to-decode-confusing-medical-bills (2026-08-02)
- Last verified: 2026-10-19

### Fill three web onboarding forms from Google Drive files with Claude in Chrome, at a heavy usage cost (partial)
- URL: https://www.realagentcases.com/use-cases/fill-three-web-onboarding-forms-from-google-drive-files-with-claude-in-chrome-at/
- Outcome: Evan Solomon asked Claude in Chrome to fill out three web onboarding forms using information scattered across their Google Drive. It worked very well: it read the pages, found the information in Drive, filled the fields and handled multi-step forms with almost no hand-holding. But it took 574 model round-trips and about 223M tokens, roughly $350-400 in API-equivalent compute, and used up the Max plan's 5-hour window in one session. Solomon notes it was slow, got slower and pricier as the session grew, and that running it on the top model was a mistake for form-filling.
- Agents: Claude in Chrome
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The forms were filled correctly but used a whole plan's usage window.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Evanzsolomon/status/2082616146734318013 (2026-07-29)
- Last verified: 2026-10-19

### Add a whole school year's calendar to the family calendar with Claude and catch clashes with work trips months ahead
- URL: https://www.realagentcases.com/use-cases/add-a-whole-school-year-s-calendar-to-the-family-calendar-with-claude-and-catch/
- Outcome: Hally Peck, a chief marketing officer in New York with two kids and a husband who also works full-time, used to schedule life only days or a week ahead. Peck gave Claude the kids' school calendars for the year, with holidays, vacations and parent-teacher conference dates, and had it add them to the family calendar and cross-check school events against existing plans. It flagged a few conflicts with work trips already booked. Catching them months in advance rather than days or weeks means the family can resolve or plan around them.
- Agents: Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/working-mom-ai-agents-handle-groceries-camp-forms-school-calendars-2026-7 (2026-07-17)
- Last verified: 2026-10-19

### Get to the ER in time for suspected testicular torsion after Claude flagged it as an emergency
- URL: https://www.realagentcases.com/use-cases/get-to-the-er-in-time-for-suspected-testicular-torsion-after-claude-flagged-it-a/
- Outcome: DANN's pain started out of nowhere on a Friday night and within twenty minutes they could barely walk; the plan was to take painkillers and sleep it off. They asked Claude first, and it said this could be testicular torsion, that there's a 4 to 6 hour window, to go to the ER now, not to eat or drink in case of surgery and not to drive themselves. They went. The first ER untwisted it and sent them home; at 3am the pain came back and after a second visit they were taken by ambulance to a bigger hospital, staying three days. Torsion was never confirmed because it was untwisted before an ultrasound, but DANN is home, healing and kept everything.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/DannPetty/status/2077064952251056195 (2026-07-14)
- Last verified: 2026-10-19

### Turn a spare ESP32 box into a baby monitor that only wakes parents after 10 minutes of fussing, built with Claude
- URL: https://www.realagentcases.com/use-cases/turn-a-spare-esp32-box-into-a-baby-monitor-that-only-wakes-parents-after-10-minu/
- Outcome: pugio and their spouse were sleep-training their baby, who wakes a few times a night and either babbles back to sleep or fusses for 10 minutes before a parent needs to go in. With a spare esp32-box-3 and a few rounds with Claude, they built a baby monitor that tracks how long the baby cries or fusses, streams audio only after 10 minutes, and sends notifications at 30 seconds and at 10 minutes. The log shows the baby often wakes 1-2 times a night and resettles almost immediately, and the parents only wake when the audio comes through. During the day it tells them when naps end.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48800997 (2026-07-06)
- Last verified: 2026-10-19

### Clear out one sender's promo emails with Gemini Spark by asking for a count before anything is deleted
- URL: https://www.realagentcases.com/use-cases/clear-out-one-sender-s-promo-emails-with-gemini-spark-by-asking-for-a-count-befo/
- Outcome: The writer behind the Stacktraced blog had months of SiriusXM trial-expiry emails from a single sender. They asked Gemini Spark to find every email from that address, report the count and date range before doing anything, then move them to Trash without touching any other sender. It found exactly 20 messages, all from that sender, and asked for confirmation. After a 'yes', it still asked for approval one email at a time, 20 separate clicks, before reporting all 20 moved to Trash. The writer checked: none left in the inbox, all 20 recoverable in Trash and nothing else disturbed, but notes native Gmail search would have been faster for a known sender.
- Agents: Gemini Spark
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.stacktraced.blog/20-emails-20-clicks-gemini-spark-cleaned-my-inbox-and-proved-why-agentic-has-limits (2026-06-28)
- Last verified: 2026-10-19

### Get a commercial cooler that kept cutting out working again for about $30 in parts, troubleshooting with Claude
- URL: https://www.realagentcases.com/use-cases/get-a-commercial-cooler-that-kept-cutting-out-working-again-for-about-30-in-part/
- Outcome: Jassy's homestead relies on a commercial glass-door cooler for garden harvest, eggs and bacon mid-cure, but it had sat broken for 8 months because a repair call would cost $600 or more. Jassy described the symptoms to Claude: it cooled for 24–48 hours, then cut out while the fan kept running, with no icing. Claude came back with six things to check, saying which were reasonable DIY jobs. It suggested cleaning the coils. After that, Jassy reported that the fan ran and they could hear it calling for the condenser but the condenser did nothing. Claude was then fairly confident it was a bad start relay or overload protector, about $15 each. Jassy replaced both themselves, the cooler cooled almost immediately and had held steady for a month. Parts were about $30 with shipping.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://thejonestead.com/how-i-saved-1200-on-appliance-repairs/ (2026-06-28)
- Last verified: 2026-10-19

### Reset a front-load washer stuck with the door locked and the pump running nonstop, with Claude and a photo of the control panel
- URL: https://www.realagentcases.com/use-cases/reset-a-front-load-washer-stuck-with-the-door-locked-and-the-pump-running-nonsto/
- Outcome: Jassy's Whirlpool Duet washer had been run without the water turned on, and got stuck in a door-fault loop: pump running, door locked, lights flashing, and no way to reset it or turn it off. Local technicians charge a $300 minimum just to show up. The first couple of reset sequences Claude suggested didn't work, but Jassy kept describing what was happening and then shared a photo of the control panel. Claude then recommended holding the cancel button for a full 3 seconds (which, Jassy notes, is printed on the panel), and the washer reset and finished its cycle. The fix cost nothing.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://thejonestead.com/how-i-saved-1200-on-appliance-repairs/ (2026-06-28)
- Last verified: 2026-10-19

### Land a new software job within a month of quitting, using ChatGPT for interview prep plans
- URL: https://www.realagentcases.com/use-cases/land-a-new-software-job-within-a-month-of-quitting-using-chatgpt-for-interview-p/
- Outcome: Rathesh Prabakar, a senior software engineer in Bengaluru, quit in October 2025 without a job lined up. Prabakar applied before work and prepared after, and asked ChatGPT to act as an experienced résumé writer. For each interview, Prabakar gave it their skills, experience and the job description plus the time available and asked for a prep plan; a high share of the likely questions it suggested came up. Questions Prabakar couldn't answer went back to ChatGPT afterwards, and doubtful answers were cross-checked on Google. With a referral from a LinkedIn contact, Prabakar accepted an offer in early November.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/quit-job-ai-linkedin-referral-2026-6 (2026-06-19)
- Last verified: 2026-10-19

### Prepare and file a property tax grievance on deadline day, between meetings, with Claude Cowork (partial)
- URL: https://www.realagentcases.com/use-cases/prepare-and-file-a-property-tax-grievance-on-deadline-day-between-meetings-with/
- Outcome: Wes Pomeroy's town assessment had nearly doubled since they bought the house eight or nine years earlier, and they realised the last day to challenge it was that day. Using Claude Cowork connected to Gmail, Google Drive and Chrome, they first asked about the official rules and deadline, which turned up that a mailed form wouldn't count: it had to be handed in in person by 9 PM. Claude found an old broker appraisal, the assessor's email about a renovation and the addition's square footage, and cut a weak argument about crime risk that had no data behind it. It filled the official form, drafted a cover letter, built a comparison page and a labelled photo exhibit and merged them into one file. Pomeroy says the first pass was about 65% right and corrected it several times. Both owners signed and the packet was stamped at the assessor's office before the deadline. The post doesn't give a result yet.
- Agents: Claude Cowork
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The grievance was filed on time; the post doesn't report the result.
- Prompt type: Original (as published in the source)
- Source: https://wespomeroy.substack.com/p/property-tax-grievance (2026-06-16)
- Last verified: 2026-10-19

### Turn rambling notes into a one-page pet-sitter guide for two dogs and a cat with Claude
- URL: https://www.realagentcases.com/use-cases/turn-rambling-notes-into-a-one-page-pet-sitter-guide-for-two-dogs-and-a-cat-with/
- Outcome: Before a trip, Chris Ivey usually stayed up the night before writing instructions for the pet sitter. This time Ivey gave Claude rambling notes about the two dogs and the cat, and Claude turned them into a clean one-page house-sitting guide covering the feeding schedule, medications, the door code and the vet's details. After a couple of tweaks it was done, and Claude also blurred the personal details so Ivey could share the finished guide publicly.
- Agents: Claude
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/christopher-ivey-344ab935_my-wife-and-i-recently-got-back-from-a-trip-activity-7471673529709572096-X-MR (2026-06-13)
- Last verified: 2026-10-19

### Shortlist cities to move to from the whole family's must-haves with ChatGPT, then visit and choose one
- URL: https://www.realagentcases.com/use-cases/shortlist-cities-to-move-to-from-the-whole-family-s-must-haves-with-chatgpt-then/
- Outcome: Before moving within Denmark, Vera Termühlen-Enger's family first talked through what each of them needed: school, nature, community, health support, job opportunities, and for their son, a skate park nearby. Only then did they ask ChatGPT to analyse possible cities against those criteria. It suggested 8 places; they crossed 3 off from what they already knew and visited the remaining 5 during the winter holidays, walking around, speaking to schools and talking to people. They chose Odense, a city they hadn't known and probably wouldn't have found without ChatGPT, though the final reasons were the school principal, the atmosphere and the neighbourhood.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/vera-term%C3%BChlen-enger-879a5150_futureofwork-leadership-decisionmaking-activity-7467812273324077056-vuqe (2026-06-03)
- Last verified: 2026-10-19

### Get out-the-door quotes from 25 dealerships at once with a ChatGPT list and message, and pay $7,000 less
- URL: https://www.realagentcases.com/use-cases/get-out-the-door-quotes-from-25-dealerships-at-once-with-a-chatgpt-list-and-mess/
- Outcome: Isaiah Crossman, a salesperson, started shopping a few days before the end of the month, when dealers chasing monthly targets are most motivated. After a test drive and a benchmark price from the nearest dealer, Crossman had ChatGPT list the 25 closest dealerships with links to their contact forms and draft one message: driven it, ready to buy, this spec and these colours, send your best out-the-door price. Sending it to all 25 took about 7 minutes. Over about 2 hours of calls across a few days, a general manager 40 minutes away made a same-day deal: $7,000 less than one dealer's earlier offer and $2,100 less than the next lowest.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/isaiahcrossman_i-bought-a-new-car-this-weekend-paid-7000-activity-7467210468965818368-RpX8 (2026-06-01)
- Last verified: 2026-10-19

### Draft an email with a couple's average monthly grocery spending from a budget spreadsheet with Gemini Spark
- URL: https://www.realagentcases.com/use-cases/draft-an-email-with-a-couple-s-average-monthly-grocery-spending-from-a-budget-sp/
- Outcome: The Verge's Jay Peters asked Gemini Spark to draft an email to their spouse with the couple's average monthly grocery spending in 2026, without naming the spouse or saying where the budget was; the spreadsheet in Drive doesn't have 'budget' in its name. Spark found the spouse's email address, pulled the monthly grocery totals from the right 2026 spreadsheet (including May's incomplete data), averaged them and put a draft in Gmail that used the spouse's first name and a sign-off the couple use. Jay's reaction: 'Wow, that's actually nuts.' Their wider verdict was that Spark still needs checking and isn't worth the $99.99-a-month plan on its own.
- Agents: Gemini Spark
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.theverge.com/tech/941138/google-gemini-spark-ai-agent-hands-on (2026-06-01)
- Last verified: 2026-10-19

### Check this week's appointments with ChatGPT's calendar access, and get one wrong by 24 hours (failure)
- URL: https://www.realagentcases.com/use-cases/check-this-week-s-appointments-with-chatgpt-s-calendar-access-and-get-one-wrong/
- Outcome: Ali van der Laan found Claude struggled with Google Calendar on an iPhone and switched to ChatGPT for calendar management. Their first question was what they had on that week. ChatGPT gave the wrong time for an appointment, off by 24 hours, and then wrote reams justifying it.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. ChatGPT gave an appointment time that was 24 hours out and defended it.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AlivanderLaan/status/2055995037507305760 (2026-05-17)
- Last verified: 2026-10-19

### Get a $22,604 ER bill in collections waived after Claude helped find grounds to challenge it
- URL: https://www.realagentcases.com/use-cases/get-a-22-604-er-bill-in-collections-waived-after-claude-helped-find-grounds-to-c/
- Outcome: Two years after an ER visit, Jackie Davalos got a collections notice saying $22,604 was owed to the hospital, having never received a bill; the records wrongly listed Davalos as uninsured. A first dispute only removed one medication charge. Davalos's partner, Walter Kerr, uploaded the billing and medical records to Claude and asked whether they had any further recourse. Claude suggested several arguments, including that legal requirements on debt and insurance might not have been met. Using many of them, Kerr wrote to executives at the hospital and its parent company, and shortly after the hospital waived the entire bill, later acknowledging a clerical error. A legal aid lawyer who reviewed it said some of Claude's analysis was right but it misunderstood some of the laws it cited.
- Agents: Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.seattletimes.com/business/people-are-using-chatbots-to-fight-medical-bills-with-mixed-results/ (2026-04-26)
- Last verified: 2026-10-19

### Check whether a half-hourly electricity tariff would be cheaper by giving Claude 87 days of smart meter readings
- URL: https://www.realagentcases.com/use-cases/check-whether-a-half-hourly-electricity-tariff-would-be-cheaper-by-giving-claude/
- Outcome: Beckie Brown's household was on a fixed Octopus Energy tariff at 26.90p per kWh and wondered about switching to Agile, whose price changes every half hour. They downloaded 87 days of half-hourly smart meter readings, Octopus's published historical Agile prices for their region and their own tariff details, and asked Claude to work it out from the real data. In about 30 minutes Claude matched usage to prices, projected a full year from three years of Agile history, and produced an 8-page report: about £171 a year saved with no changes, around £218 with small habit shifts, a peak penalty of about £133 from 4 to 7 pm, and no exit fee. They switched. In a follow-up post on 12 May 2026, the first 30-day bill was just under £78, against about £120 for the same usage on the old tariff, in a record month for negative prices.
- Agents: Claude
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://notesandsparks.substack.com/p/how-ai-helped-me-switch-energy-plans (2026-04-07)
- Last verified: 2026-10-19

### Book a monthly allergy-prescription appointment on a fiddly Japanese clinic site with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/book-a-monthly-allergy-prescription-appointment-on-a-fiddly-japanese-clinic-site/
- Outcome: Ved Kamat, who lives in Japan, gets the same pollen-allergy prescription every month and finds clinic booking sites awkward, with odd layouts and the same details asked for again and again. They had Claude in Chrome do the booking: it found the booking page, chose the return-patient appointment type, picked the earliest slot as asked and filled in the patient details. Kamat only had to type their password when Claude asked; payment details would also have stayed with them. They plan to hand over more non-critical bookings.
- Agents: Claude in Chrome
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/vedkamat_ai-booked-my-doctors-appointment-at-a-japanese-activity-7439211356303814656-3AnB (2026-03-16)
- Last verified: 2026-10-19

### Get final paychecks released within 30 minutes with an email citing state labor law, written with Grok
- URL: https://www.realagentcases.com/use-cases/get-final-paychecks-released-within-30-minutes-with-an-email-citing-state-labor/
- Outcome: After two weeks of trying to get final paychecks from a former employer, Rolo had Grok help write an email citing Illinois labor laws and the consequences for the company. Several people at the company replied within 30 minutes to say the checks would be ready for pickup. Grok was explicit that it isn't an attorney.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Rolo_1776/status/2032462097167077569 (2026-03-13)
- Last verified: 2026-10-19

### Get a stalled Affirm refund moving after 2.5 months of calls with an email drafted with Grok
- URL: https://www.realagentcases.com/use-cases/get-a-stalled-affirm-refund-moving-after-2-5-months-of-calls-with-an-email-draft/
- Outcome: Rolo had been chasing a refund from Affirm for 2.5 months, with at least 5 calls averaging 20 minutes each and no result. Grok helped draft an email that cited the relevant agencies, and within an hour Affirm emailed to say the refund check was on its way.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Same post as the final-paycheck entry; this is the second task it describes.
- Source: https://x.com/Rolo_1776/status/2032462097167077569 (2026-03-13)
- Last verified: 2026-10-19

### Estimate the year's federal, state and local tax still due from a folder of tax documents with Claude Cowork, before the accountant reports
- URL: https://www.realagentcases.com/use-cases/estimate-the-year-s-federal-state-and-local-tax-still-due-from-a-folder-of-tax-d/
- Outcome: Anthony Garcia's taxes had got complicated enough that Garcia hired an accountant, who was still a month and a half away from giving the full picture. Garcia copied all the tax files, some electronic PDFs and some scans of paper copies, into one folder, pointed Claude Cowork at it and asked for an estimate of the 2025 taxes due. Claude read all of them. In 30 minutes it produced a summary table of tax liability, amount withheld and balance due, with a thorough breakdown that referred to boxes on the tax forms.
- Agents: Claude Cowork
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://acgarcia21.substack.com/p/claude-cowork-can-do-your-taxes (2026-02-22)
- Last verified: 2026-10-19

### Revise for a school exam without the textbook by giving ChatGPT the official PDF to answer the exercises
- URL: https://www.realagentcases.com/use-cases/revise-for-a-school-exam-without-the-textbook-by-giving-chatgpt-the-official-pdf/
- Outcome: SriSakthi Saravana Prabhu's younger child had a Kannada exam the next day after a week off school with fever, and the textbook was likely left at school. The parent can speak Kannada but can't read or write it, and the exam questions come from the exercises at the end of each lesson. They found the Grade 3 textbook as a PDF on the Karnataka education portal and uploaded it to ChatGPT. Within seconds it gave answers, and the child confirmed they were the ones needed. Some pages came back without answers, but uploading those pages separately filled the gaps.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/sakthishanmugam_aiineducation-learningneverstops-parenting-activity-7426615892194566144-2lAx (2026-02-09)
- Last verified: 2026-10-19

### Let Claude Cowork tidy a family desktop, and lose 15 years of camera photos until iCloud recovery brings them back
- URL: https://www.realagentcases.com/use-cases/let-claude-cowork-tidy-a-family-desktop-and-lose-15-years-of-camera-photos-until/
- Outcome: Nick Davidov asked Claude Cowork to organise their spouse's desktop. It asked permission to delete temporary Office files, they agreed, and then it said "ooops": while renaming, it had deleted a folder with every photo the spouse had taken on their camera over 15 years. The deletion was done in the terminal, so nothing was in the Trash; iCloud had already synced the new structure, there was no Time Machine backup, and disk recovery tools found nothing. Apple pointed them to iCloud's recovery feature, which keeps removed iCloud Drive files for 30 days, and the photos started coming back. Their advice: don't let these tools near anything hard to repair or not under version control.
- Agents: Claude Cowork
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The agent deleted 15 years of photos; they came back through iCloud's recovery feature, not the agent.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ndavidov_asked-claude-cowork-organize-my-wifes-desktop-activity-7426055608010952704-ghzL (2026-02-08)
- Last verified: 2026-10-19

### Win a rent tribunal complaint over a move-out repainting bill with ChatGPT structuring the evidence
- URL: https://www.realagentcases.com/use-cases/win-a-rent-tribunal-complaint-over-a-move-out-repainting-bill-with-chatgpt-struc/
- Outcome: After moving out of a rental flat in Denmark after one year, Andreas Strandfelt was billed a surprisingly high amount for repainting, despite detailed move-in documentation, a thorough move-out inspection and evidence that an identical neighbouring flat had been painted more for less that month. The landlord wouldn't explain the costs. Strandfelt used ChatGPT to structure a complaint to the local rent tribunal, referencing the email thread, the move-in and move-out reports and the relevant legal paragraphs. The tribunal ruled in their favour, and the landlord had to return more of the deposit. The costs were 300 DKK to file and less than 200 DKK for a month of ChatGPT Plus.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/strandfelt_i-won-a-case-against-my-landlord-with-help-activity-7419319626062258176-VzZ9 (2026-01-20)
- Last verified: 2026-10-19

### Spot when an escaped indoor cat comes back to the door from days of camera footage with Gemini
- URL: https://www.realagentcases.com/use-cases/spot-when-an-escaped-indoor-cat-comes-back-to-the-door-from-days-of-camera-foota/
- Outcome: Zachary Thorman's skittish indoor cat, Puff, had slipped out 72 hours earlier without anyone noticing, with nights around 30°F. Thorman had Gemini go through days of Nest camera history. It found a pattern: Puff was coming back to check the door during quiet hours, around 7 AM and after 10 PM. Gemini suggested waiting rather than searching, laying a tuna-broth scent trail and leaving the door cracked; about an hour into waiting, Puff ran inside. Thorman then used Gemini to find local vet availability and book a check-up.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/zthor_gemini-googlecloud-appliedai-activity-7417193272860651521-8QYj (2026-01-14)
- Last verified: 2026-10-19

### Turn two weeks of paper classroom surveys into a data sheet, report and parent email with Gemini
- URL: https://www.realagentcases.com/use-cases/turn-two-weeks-of-paper-classroom-surveys-into-a-data-sheet-report-and-parent-em/
- Outcome: Jessica McKellar volunteers as chair of the parent advisory committee at their daughter's two-campus preschool. When the school changed food programmes and got a lot of parent feedback, it ran two weeks of classroom surveys tracked on paper. McKellar photographed the sheets and gave them to Gemini, which turned them into spreadsheets of raw data, analysed and summarised them in a management report for the administrative team, and drafted an email for the school to send parents explaining the results.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/jesstess_i-took-inventory-of-how-i-have-used-ai-for-activity-7417304383396167680-pzVo (2026-01-14)
- Last verified: 2026-10-19

### Fight a private parking charge with ChatGPT finding holes in the operator's evidence, then recover £462.15 in costs
- URL: https://www.realagentcases.com/use-cases/fight-a-private-parking-charge-with-chatgpt-finding-holes-in-the-operator-s-evid/
- Outcome: Loren Cole, a freelance chef, got a £60 parking charge for overstaying by five minutes at a retail park, despite paying for a three-hour ticket after the payment machines played up. When the operator threatened court, Cole used ChatGPT to scrutinise its letters and evidence: it pointed out claims of non-payment despite a successful card payment, a floor plan sent instead of photos of the signs, and missing detail on entry and exit times against the 10-minute grace period, and it drafted replies. Cole added photos showing the signage didn't match the floor plan, and the operator discontinued the claim. Cole then used ChatGPT to claim litigant-in-person costs at £19 per hour; after a default costs certificate and a ChatGPT-drafted final warning, the operator paid £462.15.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.mirror.co.uk/news/uk-news/i-used-chatgpt-fight-parking-36429227 (2025-12-19)
- Last verified: 2026-10-19

### Plan a nut-free family Christmas dinner that bakes in one oven at one temperature with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-nut-free-family-christmas-dinner-that-bakes-in-one-oven-at-one-temperatur/
- Outcome: Stephanie M. Morrison was hosting family and asked ChatGPT for a menu with a list of criteria: easy, with some store-bought items and a few simple homemade dishes; dishes that could bake at the same time and temperature in a single oven; no nuts, red meat or anything spicy; traditional enough for a crowd; and something picky kids would try. From several options, Morrison chose roasted orange and cranberry chicken with roasted red potatoes with garlic and parmesan, plus store-bought rolls, salad and dessert. ChatGPT said what to start when so everything finished together, and gave a shopping list. It took 5 minutes, and Morrison says it worked perfectly and was a hit.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/stephanie-m-morrison_i-had-chatgpt-plan-christmas-dinner-we-activity-7406340742593544193-hqgL (2025-12-15)
- Last verified: 2026-10-19

### Pick a family health plan at open enrollment from photos of the benefits booklet with ChatGPT, then cross-check the numbers
- URL: https://www.realagentcases.com/use-cases/pick-a-family-health-plan-at-open-enrollment-from-photos-of-the-benefits-booklet/
- Outcome: Chris L.'s family spent about $19k on healthcare in a year that wasn't especially bad health-wise, because they'd hit the out-of-pocket maximum on a plan that didn't fit their needs. During open enrollment they photographed the pricing and coverage pages of the enrollment booklet and gave them to ChatGPT, with the care they expected. After 5 and a half minutes it compared the PPO, Gold and Bronze plans, explained how an FSA and an HSA would work out, and gave best- and worst-case yearly costs for each. They then asked about the costs in between those two extremes and got another detailed answer. They checked the output with two other models, which gave the same numbers, and expect to pay between $2,600 and $16,500 in 2026, saving at least $2,500 compared with the year before. It took them under 20 minutes.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/cwlindeman_open-enrollment-closes-today-dont-understand-activity-7395107290720247808-712P (2025-11-14)
- Last verified: 2026-10-19

### Build a custom GPT as a job-search coach and go from 2 or 3 interviews to at least 7 and a PayPal offer
- URL: https://www.realagentcases.com/use-cases/build-a-custom-gpt-as-a-job-search-coach-and-go-from-2-or-3-interviews-to-at-lea/
- Outcome: Product manager Amar Saurabh, ex-Meta and TikTok, got only two or three interviews in the first two months of applying, and plain ChatGPT kept giving generic answers and asking for the résumé again. In about an hour and a half Saurabh built a custom GPT, 'PM job search advisor', loaded with their résumé, LinkedIn profile and project details. It drafted recruiter outreach (including 300-character connection requests), tailored the résumé to each job description and prepped each interview. Over two months Saurabh landed at least 7 interviews, with companies including Reddit, Intuit and PayPal, and took a lead product manager job at PayPal.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/how-to-use-custom-gpt-make-job-search-easier-prompts-2025-10 (2025-10-25)
- Last verified: 2026-10-19

### Compare two contractor quotes for a driveway and retaining wall with ChatGPT and catch the missing drainage
- URL: https://www.realagentcases.com/use-cases/compare-two-contractor-quotes-for-a-driveway-and-retaining-wall-with-chatgpt-and/
- Outcome: Jeffrey Trull had two quotes for a driveway extension and retaining wall worth a few thousand dollars, from contractors they'd never used. One was $500 cheaper. They fed both into ChatGPT, which took 5 minutes. It flagged that the cheaper quote didn't mention drainage or backfill, and explained that water building up behind a retaining wall can make it fail; the dearer quote included a full drainage system with stone backfill. It also set out the trade-off between the nicer interlocking blocks and basic cement blocks, and wrote follow-up questions for each contractor, since one quote gave no wall dimensions and the other didn't explain ground prep. Trull planned to take those questions back to both contractors, and to ask the cheaper one to add drainage and backfill, before signing anything.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. Useful analysis and questions; the post was written before a contractor was chosen.
- Prompt type: Original (as published in the source)
- Source: https://moneymeetsai.com/p/ai-contractor-quotes-saved-mistake (2025-10-15)
- Last verified: 2026-10-19

### Find a severed transmission cooler line from a roadside call with Grok, then work out a hose fix that held
- URL: https://www.realagentcases.com/use-cases/find-a-severed-transmission-cooler-line-from-a-roadside-call-with-grok-then-work/
- Outcome: Shan Newton was hauling a heavy load up a mountain when a huge cloud of white smoke appeared behind the truck and fluid leaked down the road. Knowing nothing about fixing cars, Newton described what they saw to Grok 4 by voice; it said the white smoke pointed to the transmission and where to look, and the transmission oil cooler line had completely severed. After a passer-by towed the truck 18 miles into town, the new hose wouldn't fit inside the factory-crimped metal pipe. Grok suggested a hose with a bigger inner diameter to fit over the pipe, estimated the pressure at 10 - 20 PSI and said metal hose clamps would be enough. The workaround worked, and Newton says the truck runs better than before.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/shannewton_i-was-hauling-a-heavy-load-up-a-mountain-activity-7361835806967123969-owVR (2025-08-14)
- Last verified: 2026-10-19

### Cut broadband and energy bills by about £782 a year with negotiation scripts from ChatGPT (partial)
- URL: https://www.realagentcases.com/use-cases/cut-broadband-and-energy-bills-by-about-782-a-year-with-negotiation-scripts-from/
- Outcome: Finance writer Jessie Hewitson tested ChatGPT on household bills. It had Hewitson find a cheaper competing broadband deal (£23 a month with Virgin against £35 with Vodafone) and wrote a script; read word for word on the phone, it got Vodafone to offer a £22-a-month reduction, which Hewitson says saved £156 a year. An emailed tariff review with Octopus turned into a phone call that moved them to tariffs worth an estimated £626 a year. Not everything worked: a ChatGPT-written appeal against a £90 Transport for London penalty failed, and a home-insurance quote it led to came in at £1,627.24 against £436 before.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The penalty appeal ChatGPT wrote was rejected, and the insurance quote it led to was far higher than the old policy.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://inews.co.uk/inews-lifestyle/chat-gpt-cut-household-bills-782-richer-3820406 (2025-07-25)
- Last verified: 2026-10-19

### Stop an insurer's repeated document requests with a firm ChatGPT letter and get ₹85,000 approved the next day
- URL: https://www.realagentcases.com/use-cases/stop-an-insurer-s-repeated-document-requests-with-a-firm-chatgpt-letter-and-get/
- Outcome: Jacinth Paul Z submitted a ₹89,000 hospitalisation claim, of which ₹65,000 was marked claimable, and sent every required document within a week. After three months of silence the insurer asked for more documents three times, each needing a trip to the hospital. The fourth time, Jacinth used ChatGPT to draft a firm, legal-toned letter saying enough documents had already been submitted, that the requests looked like a delay tactic, that no more would be sent, and that HR was being told how the claim had been handled. Jacinth uploaded the letter instead of more documents, and the claim was approved the next day for ₹85,000.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/jacinthp_ai-chatgpt-insurance-activity-7348545532170031104-2-0_ (2025-07-09)
- Last verified: 2026-10-19

### Build a goodwill case with ChatGPT and get a €6,000 out-of-warranty engine replacement for free
- URL: https://www.realagentcases.com/use-cases/build-a-goodwill-case-with-chatgpt-and-get-a-6-000-out-of-warranty-engine-replac/
- Outcome: Thomas Weilbach's car was out of warranty, and the dealership said the only long-term fix, a full engine replacement quoted at €6,000, would be paid out of pocket. Weilbach used ChatGPT in research mode to dig into the symptoms (oil consumption, juddering under load and limp mode), known problems with that engine, service bulletins and how others had handled similar cases, testing each possible cause against the symptoms. ChatGPT then helped turn it all into a clear, structured argument. Weilbach took the case to the dealership, and the manufacturer approved the full engine replacement under goodwill at no charge.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/thomas-weilbach_how-ai-assisted-search-saved-me-5978-ive-activity-7343383984359325696-Cjru (2025-06-24)
- Last verified: 2026-10-19

### Cook a week of meals for one from ingredients about to spoil with a Gemini recipe for under $12
- URL: https://www.realagentcases.com/use-cases/cook-a-week-of-meals-for-one-from-ingredients-about-to-spoil-with-a-gemini-recip/
- Outcome: Daniel G. Taylor, who is single and wants to cook once for the week, had beef chunks near their use-by date, soft carrots, sprouting potatoes and half an onion. Taylor listed the ingredients for Gemini with the constraints: feed one person for multiple meals, use pantry staples and need minimal extra shopping. Gemini produced a beef and winter vegetable stew that cost under $12, will feed Taylor for five days and used up food that was heading for the bin.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/daniel-g-taylor-9730654_ai-practicaltips-foodwaste-activity-7340358677821911041-K9Ar (2025-06-16)
- Last verified: 2026-10-19

### Cut a family's grocery spend from $400–$500 every few days to $140 a week with a ChatGPT meal plan
- URL: https://www.realagentcases.com/use-cases/cut-a-family-s-grocery-spend-from-400-500-every-few-days-to-140-a-week-with-a-ch/
- Outcome: When Brooke Ferrier, a parent of two on the NSW Central Coast, and their partner both lost their jobs, Ferrier asked ChatGPT for a low-cost meal plan for two adults and two toddlers, then refined it with follow-ups: hidden vegetables, low-carb options and no lentils. ChatGPT gave one recipe per meal with variations for each person, compared supermarket prices and produced a categorised shopping list with quantities. A week of breakfasts, dinners and snacks came to $140, against the $400–$500 Ferrier had been spending every four or five days, with much less food thrown out. Ferrier also uses it to find recipes for long-forgotten pantry items.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.abc.net.au/news/2025-05-30/using-artificial-intelligence-to-plan-meals-save-grocery-money/105302996 (2025-05-30)
- Last verified: 2026-10-19

### Get a reluctant reader through spring-break reading homework with ChatGPT stories about their favourite topic
- URL: https://www.realagentcases.com/use-cases/get-a-reluctant-reader-through-spring-break-reading-homework-with-chatgpt-storie/
- Outcome: Lisa Durant's almost-9-year-old has special needs, struggles with reading and resists books without pictures, so reading homework often ended in meltdowns. For the spring-break assignment (20 minutes of reading at least five times, any material), Durant asked ChatGPT for a short, silly, age-appropriate story about tornadoes, the child's special interest, starring the child and a younger sibling, and had the child read it aloud. The child soon started writing their own prompts for new stories, and reads them without protest, picking up good vocabulary along the way.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/lisa-durant-17468540_how-chatgpt-is-helping-my-special-needs-daughter-activity-7319698938339487744-t6ta (2025-04-20)
- Last verified: 2026-10-19

### Find and book a same-day packing-and-moving company for a two-destination move with ChatGPT Deep Research
- URL: https://www.realagentcases.com/use-cases/find-and-book-a-same-day-packing-and-moving-company-for-a-two-destination-move-w/
- Outcome: Nicole Leffer had a complicated move the following week: belongings going to two different locations, the busiest work month ever and sciatica that made packing painful. Leffer gave ChatGPT Deep Research the from and to addresses, the schedule and the need for a same-day mover that could pack and transport in one day, and asked it to check reviews and track record. Within minutes it presented the best options. Leffer checked that the top pick really had the 1600+ 5-star Google reviews it claimed, and fifteen minutes after entering the prompt was on the phone booking. Leffer says the movers were perfect, with fair pricing.
- Agents: ChatGPT Deep Research
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/nicoleleffer_chatgpt-deep-research-and-google-deep-research-activity-7318664604337725441-E1y4 (2025-04-17)
- Last verified: 2026-10-19

### Turn long recipe steps into an importable to-do list with a custom GPT so every Thanksgiving dish is on time
- URL: https://www.realagentcases.com/use-cases/turn-long-recipe-steps-into-an-importable-to-do-list-with-a-custom-gpt-so-every/
- Outcome: Amanda Parsons, who has adult ADHD, finds recipes pack too much into one step. While planning Thanksgiving, Parsons asked ChatGPT for help with the cooking schedule, then found it could turn recipes into CSV files for the Todoist to-do app; for the first time, dinner was ready on time and every dish made it to the table as planned. After a chaotic batch of English muffins without it, Parsons built a custom GPT to do the conversion, going through five rounds of tests over an hour until the CSV imported successfully.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://goodworkpeople.substack.com/p/good-work-people-tame-kitchen-chaos-with-chatgpt (2025-02-04)
- Last verified: 2026-10-19

### Make a grade-4 math practice game with ChatGPT after its first try didn't work
- URL: https://www.realagentcases.com/use-cases/make-a-grade-4-math-practice-game-with-chatgpt-after-its-first-try-didn-t-work/
- Outcome: The poster asked ChatGPT for a math practice worksheet for their fourth-grade daughter. Its first attempt, an HTML page, didn't work. It then rendered a personalised practice game directly in the chat, which worked perfectly; the daughter's verdict was 'beat that IXL!', the homework software her school assigns.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/rida/status/2110521012043317606 (2026-10-14)
- Last verified: 2026-10-18

### Get a free replacement toothbrush handle out of warranty by having Grok Bot work customer service
- URL: https://www.realagentcases.com/use-cases/get-a-free-replacement-toothbrush-handle-out-of-warranty-by-having-grok-bot-work/
- Outcome: Over a week, the poster's Grok Bot, connected to the Bland AI phone service, dealt with Philips customer service by phone and chat about an out-of-warranty Sonicare toothbrush with a dead battery. Along the way it got two 30%-off coupon codes and free brush heads, and in the end a free replacement handle, which the poster values at about $260. The poster says they never spoke to customer service themselves. It also reviewed their PG&E electricity use and filled out forms to switch plans to suit charging two EVs.
- Agents: Grok Bot
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/Jonathan_Miras/status/2109032471287456086 (2026-10-10)
- Last verified: 2026-10-18

### Track where you stored passports, documents and spare keys during a deep clean with Gemini and Find Hub
- URL: https://www.realagentcases.com/use-cases/track-where-you-stored-passports-documents-and-spare-keys-during-a-deep-clean-wi/
- Outcome: Android Police writer Akshay Bhalla tends to forget where things are put away, which made the yearly deep clean drag. This year Bhalla used Gemini voice commands to save items to Find Hub's 'Remembered' section as they went: passport, tax documents, birth certificates, spare house and car keys, even a tape measure, each with a photo of where it was stored. Asking Gemini 'Where did I put my passport?' brings back the spot and the picture. The cleaning took two hours instead of a full day, though items have to be updated by hand when moved.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Uses Find Hub's Remembered feature with Gemini on Android 16 or newer.
- Source: https://www.androidpolice.com/gemini-find-hub-remembered-items/ (2026-10-08)
- Last verified: 2026-10-18

### Plan 15 family meals a week from a recipe library and make the shopping list with Grok Bot
- URL: https://www.realagentcases.com/use-cases/plan-15-family-meals-a-week-from-a-recipe-library-and-make-the-shopping-list-wit/
- Outcome: After a decade of trying to automate meal planning (a spreadsheet, then a self-built app), the poster now has a custom Grok Bot that plans 15 meals a week for the family. It keeps a library of recipes as Markdown files, makes a printable weekly meal plan and produces a shopping list, which the poster pastes into ChatGPT to fill an Instacart order.
- Agents: Grok Bot, ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster is chief of staff at Wispr Flow and shared a public template for the bot.
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/lyndallschreinr/status/2107888353630908437 (2026-10-07)
- Last verified: 2026-10-18

### Trace a warm fridge to a failed freezer fan and order a C$32 part with Grok
- URL: https://www.realagentcases.com/use-cases/trace-a-warm-fridge-to-a-failed-freezer-fan-and-order-a-c-32-part-with-grok/
- Outcome: When the bottom section of their fridge warmed to about 11–14°C, the poster worked through a process of elimination with Grok, which pointed to the freezer fan and found the cheapest well-reviewed replacement (C$32.18 with taxes). The fan arrived the next day; the poster says it was indeed the problem and the fridge is working perfectly now.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kjehiel/status/2103166160241852550 (2026-09-24)
- Last verified: 2026-10-18

### Make a 21-page vacation activity book and gamified chore charts for three kids with ChatGPT
- URL: https://www.realagentcases.com/use-cases/make-a-21-page-vacation-activity-book-and-gamified-chore-charts-for-three-kids-w/
- Outcome: Sara Sprong, a parent of three in Colorado, uses ChatGPT as a brainstorming partner to make printable resources. For a trip to Rhinelander, Wisconsin, it helped create a 21-page activity book for their kids and cousins, with Mad Libs about their activities and a scavenger hunt around the resort, formatted for printing. Sprong also made a 'Bedroom Boss Battle' chore chart in timed levels and a tracker for one-on-one parent–child dates, and says it has helped cut screen time at home.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: As-told-to essay: the account is the parent's, written up by the reporter.
- Source: https://www.businessinsider.com/mom-ai-activity-books-chore-charts-chatgpt-2026-9 (2026-09-22)
- Last verified: 2026-10-18

### Find a second YouTube Premium you've paid for unused for a year and a half with Grok
- URL: https://www.realagentcases.com/use-cases/find-a-second-youtube-premium-you-ve-paid-for-unused-for-a-year-and-a-half-with/
- Outcome: While getting Grok's help to tidy up too many old Google and YouTube accounts, X user Tesla Patriot found they had been paying for a second YouTube Premium account they weren't using for a year and a half. They stopped the $22-a-month charge, and Grok pointed them to calling Apple as the way to ask for a refund.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A similar entry on this site uses Gemini to spot duplicate streaming plans.
- Source: https://x.com/TeslaPatriot/status/2101734539991076944 (2026-09-20)
- Last verified: 2026-10-18

### Spot low brake fluid behind recurring brake warning lights with ChatGPT after two garage visits
- URL: https://www.realagentcases.com/use-cases/spot-low-brake-fluid-behind-recurring-brake-warning-lights-with-chatgpt-after-tw/
- Outcome: The brake warning lights on the poster's car kept coming on. A mechanic claimed to have fixed it, the lights came back, and on a second visit the garage said it couldn't work it out without a fault code. At home, the poster noticed the brake fluid was low, confirmed it with ChatGPT from a photo, bought fluid and topped it up, and says that fixed it.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Low brake fluid can be a sign of worn pads or a leak; topping up doesn't rule those out.
- Source: https://x.com/bradgessler/status/2097467852953669791 (2026-09-08)
- Last verified: 2026-10-18

### Try to book an eye appointment through Gemini's Zocdoc connection and end up booking on the clinic site
- URL: https://www.realagentcases.com/use-cases/try-to-book-an-eye-appointment-through-gemini-s-zocdoc-connection-and-end-up-boo/
- Outcome: Carrie Liken asked Gemini to book a yearly eye appointment at their usual optician. It didn't work at first because the Zocdoc app needed switching on in settings and calling with '@zocdoc'. Once it was on, Gemini offered times for only one of the practice's doctors, not Liken's. Google's business listing showed the same single doctor, so Liken went to the practice website, found their doctor listed and booked there directly.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Gemini only offered one doctor at the practice, not the writer's; the appointment was booked directly on the practice website.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The writer is a healthcare marketing consultant and uses the experience to make a point about provider data.
- Source: https://carrieliken.substack.com/p/geminis-zocdoc-integration-falls (2026-09-05)
- Last verified: 2026-10-18

### Turn every household manual into a searchable support desk with NotebookLM
- URL: https://www.realagentcases.com/use-cases/turn-every-household-manual-into-a-searchable-support-desk-with-notebooklm/
- Outcome: XDA's Megan Ellis added PDF manuals for devices around the home (TV, air purifier, dishwasher, washing machine, tumble dryer, earphones and more) to NotebookLM, now called Gemini Notebook, mostly by pasting PDF links. Asked why dishes were still dirty after a wash cycle, it gave the answers Ellis had found in the manual in a few seconds, against 30 minutes spent searching the paper manual before. It couldn't say which of two headphone buttons was power, and its description of the dryer's heat exchanger still needed checking against the manual.
- Agents: NotebookLM
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. It found the dishwasher answer quickly, but couldn't pinpoint a headphone button and its dryer part description needed checking against the manual.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.xda-developers.com/using-manuals-in-notebooklm-gemini-notebook-as-support-desk/ (2026-08-09)
- Last verified: 2026-10-18

### Fill in a GP online consultation form clearly and get seen in five minutes with Grok
- URL: https://www.realagentcases.com/use-cases/fill-in-a-gp-online-consultation-form-clearly-and-get-seen-in-five-minutes-with/
- Outcome: When home care didn't clear up their symptoms, X user David Lermarl had Grok help write the online request form for their GP practice. They say the appointment took less than five minutes, they came away with the medicine they needed, and the doctor thought the approach was great, while noting it may not work every time.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/DavidLermarl/status/2084557383146090909 (2026-08-04)
- Last verified: 2026-10-18

### Get a laptop that won't boot after a dead battery working again step by step with Grok
- URL: https://www.realagentcases.com/use-cases/get-a-laptop-that-won-t-boot-after-a-dead-battery-working-again-step-by-step-wit/
- Outcome: After their teenager let a Dell laptop's battery die, it stopped booting. The owner, who says they know nothing about IT, sent Grok a screenshot of each screen and asked what to do next. They worked through BIOS changes, Dell's recovery tools and repair commands, and finally booted from a Windows 11 installation USB, entered the BitLocker recovery key and got into recovery. The laptop now boots normally with all files intact. At the end, Grok wrote a summary of what was wrong, what was tried and what worked, which they posted.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ProfUnfiltered/status/2076323807401214255 (2026-07-12)
- Last verified: 2026-10-18

### Audit a year of medical bills against your insurance contract with ChatGPT (partial)
- URL: https://www.realagentcases.com/use-cases/audit-a-year-of-medical-bills-against-your-insurance-contract-with-chatgpt-parti/
- Outcome: CNET's Robin Hartill, a certified financial planner, hit a $10,150 out-of-pocket maximum in under five months after two eye surgeries. Hartill redacted and uploaded 14 itemised bills and EOBs, a 149-page insurance contract and a spreadsheet of 87 claims to ChatGPT with an auditor prompt. It flagged nothing new on its own and showed an insurance rep had been wrong about a $250 co-pay. When asked directly why Hartill had paid $11,512, it confirmed a $1,512 overpayment on the second surgery and explained why.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT confirmed and explained the $1,512 overpayment but didn't find it unprompted; Hartill spotted it from their own payment records, and the refund was still pending when the article was published.
- Prompt type: Original (as published in the source)
- Source: https://www.cnet.com/tech/services-and-software/chatgpt-medical-bill-errors/ (2026-07-09)
- Last verified: 2026-10-18

### Size a portable solar battery for an outdoor kitchen and pizza night with Grok
- URL: https://www.realagentcases.com/use-cases/size-a-portable-solar-battery-for-an-outdoor-kitchen-and-pizza-night-with-grok/
- Outcome: An X user asked Grok to help work out their power needs for an outdoor kitchen running off a Jackery Explorer 2000 solar battery. On the night they ran lights and hosted a pizza party with a wood-fired pizza oven, cooking six pizzas, and posted screenshots of the charge left with several lights and an Alexa still running. Their verdict: 'Grok told me right!'
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/GillelandKristi/status/2072867221924192577 (2026-07-03)
- Last verified: 2026-10-18

### Let ChatGPT handle airline and car-rental disputes from Gmail and recover about €200 so far
- URL: https://www.realagentcases.com/use-cases/let-chatgpt-handle-airline-and-car-rental-disputes-from-gmail-and-recover-about/
- Outcome: The poster connected ChatGPT to Gmail and lets it handle disputes on its own. So far it has got €100 from an airline for a broken stroller and a €108 refund from a car-rental company for a very dirty child seat; the poster puts the total at €200 so far, with a trickier dispute still ongoing.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Two claims were paid; a third, trickier one was still ongoing at the time of posting.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/haltakov/status/2069797877673513084 (2026-06-24)
- Last verified: 2026-10-18

### Clear over 9,400 unread emails safely by building a 'safe list' first with ChatGPT
- URL: https://www.realagentcases.com/use-cases/clear-over-9-400-unread-emails-safely-by-building-a-safe-list-first-with-chatgpt/
- Outcome: Back from a week away, Abha Malpani Naismith had more than 9,400 unread Gmail messages. Claude couldn't act inside the connected inbox for them, so they switched to ChatGPT with Gmail connected. It analysed the inbox by sender, subject and category without opening emails, built a 'Safe List' of financial, legal, medical, school and travel mail never to touch without approval, and then helped archive, label, delete or unsubscribe in bulk. They archived or deleted more than 4,500 emails and reached inbox zero in about 30 minutes.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The post's headline gives 9,476 unread emails; the body says 9,467.
- Source: https://abhamalpaninaismith.substack.com/p/9476-emails-to-0-in-30-minutes (2026-06-15)
- Last verified: 2026-10-18

### Turn four messy contractor bids into an apples-to-apples comparison with Claude
- URL: https://www.realagentcases.com/use-cases/turn-four-messy-contractor-bids-into-an-apples-to-apples-comparison-with-claude/
- Outcome: A Hacker News commenter had four home-repair bids with wildly different line items, costs, timelines, scopes and formats. Claude (Opus) turned them into an apples-to-apples comparison, filled gaps with reasonable inferences from the other bids, produced a PDF they printed to go through with their partner, and suggested key questions for follow-up calls. The commenter says it clearly showed the advantages of one of the bids.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48434979 (2026-06-07)
- Last verified: 2026-10-18

### Match a 1999 wood floor stain from a sunlit photo and order a quart with ChatGPT
- URL: https://www.realagentcases.com/use-cases/match-a-1999-wood-floor-stain-from-a-sunlit-photo-and-order-a-quart-with-chatgpt/
- Outcome: A Hacker News commenter needed a stain to match the wood flooring in a house built in 1999. They uploaded a clear photo taken in bright sunlight, ChatGPT searched online for a matching stain colour and presented ordering options, and a quart was delivered the day before they posted.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48426560 (2026-06-06)
- Last verified: 2026-10-18

### Find why a child's science-fair electromagnet wasn't working on a Gemini Live video call
- URL: https://www.realagentcases.com/use-cases/find-why-a-child-s-science-fair-electromagnet-wasn-t-working-on-a-gemini-live-vi/
- Outcome: A Hacker News commenter and their 10-year-old couldn't get an electromagnet experiment for the school science fair to work. On a Gemini Live video call, Gemini went through the possible causes and spotted that the insulation hadn't been scraped off properly at the end of the wire connected to the battery, which the parent couldn't see. The apparatus then worked as expected, and the commenter says the child and two friends were impressed.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48419364 (2026-06-05)
- Last verified: 2026-10-18

### Make an endless maze-puzzle generator from a puzzle your child loves with Gemini
- URL: https://www.realagentcases.com/use-cases/make-an-endless-maze-puzzle-generator-from-a-puzzle-your-child-loves-with-gemini/
- Outcome: A Hacker News commenter's daughter had finished every printable version of a favourite maths maze puzzle they could find online. The parent uploaded a picture of the puzzle to Gemini and asked for a website that generates random ones; in under a minute it produced a working generator. Gemini then added the child's ideas (more operations, larger grids), and they went on to build generators for other puzzle ideas on the spot. The commenter admits the code wasn't pretty but says it did the job.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48419696 (2026-06-05)
- Last verified: 2026-10-18

### Compare options for towing a car cross-country with your truck with ChatGPT
- URL: https://www.realagentcases.com/use-cases/compare-options-for-towing-a-car-cross-country-with-your-truck-with-chatgpt/
- Outcome: A Hacker News commenter needed to tow their daughter's car across the country behind their truck: trailer or dolly, what each costs, and whether the truck could handle the combined weight. After a couple of piecemeal questions they gave ChatGPT the whole picture in one prompt (car make, truck model, start and end points), and it came back with a report with comparison tables and calculated towing weights for the different options.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The commenter quotes the prompt with placeholders ('make A', 'truck model B', 'point C to point D'), so this is a recreated version.
- Source: https://news.ycombinator.com/item?id=48418620 (2026-06-05)
- Last verified: 2026-10-18

### Cook and freeze a vegan red lentil and potato curry from a ChatGPT recipe
- URL: https://www.realagentcases.com/use-cases/cook-and-freeze-a-vegan-red-lentil-and-potato-curry-from-a-chatgpt-recipe/
- Outcome: The poster asked ChatGPT for an authentic vegan Indian curry with potatoes, red lentils, coconut milk and tomatoes, followed the recipe with a few changes of their own (extra spices, red curry paste, ground cumin in place of seeds), and calls it the best Indian curry. They froze the leftovers in meal-prep containers with rice; the poster says they reheated well over several weeks and the family ate every one.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster runs AI marketing services for local businesses.
- Source: https://x.com/VeganReignmaker/status/2062225729757012443 (2026-06-03)
- Last verified: 2026-10-18

### Learn what limp mode is at the roadside and get moving again with ChatGPT
- URL: https://www.realagentcases.com/use-cases/learn-what-limp-mode-is-at-the-roadside-and-get-moving-again-with-chatgpt/
- Outcome: On the way to work, the poster's car suddenly crawled at about 4 mph with the check-engine light blinking and other dashboard lights on. They pulled over and asked ChatGPT, which explained limp mode. The poster switched the car off and on again, and says it worked fine afterwards, saving a panicked call to a mechanic.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A flashing check-engine light can signal a serious fault; a restart that clears limp mode doesn't mean the cause is fixed.
- Source: https://x.com/AlanRotramel/status/2057173950552273264 (2026-05-20)
- Last verified: 2026-10-18

### Find a water heater's failing check valve after other checks with Grok
- URL: https://www.realagentcases.com/use-cases/find-a-water-heater-s-failing-check-valve-after-other-checks-with-grok/
- Outcome: After flushing the water heater, cleaning filters and checking error codes and sensors, the poster wasn't sure what to check next and asked Grok, which said that model has a check valve that often fails. Pulling it out solved the problem; the poster says it was an $8 part and saved $800 in repairs.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Another case on this site also traces a water heater fault with Grok; this one comes after the owner had already done the standard checks.
- Source: https://x.com/jasonbcox0/status/2054629235202167084 (2026-05-13)
- Last verified: 2026-10-18

### Write a complaint that got a health insurer to reverse an ADD assessment refusal with ChatGPT
- URL: https://www.realagentcases.com/use-cases/write-a-complaint-that-got-a-health-insurer-to-reverse-an-add-assessment-refusal/
- Outcome: Natja Vinkler, 21, was refused cover by their Danish health insurer, PFA, for a private ADD assessment because they'd been assessed as a child. Vinkler gave ChatGPT the correspondence and their childhood records, then spent several days going back and forth, adding PFA's guidelines, asking for different drafts and fact-checking everything before their mother and a colleague read it. About a month after Vinkler sent it, PFA called to say it would cover the assessment. PFA confirmed to TV 2 that the complaint led it to look again.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: PFA told TV 2 its first decision was based on incomplete childhood records, and the complaint brought the full records to its attention.
- Source: https://nyheder.tv2.dk/samfund/2026-05-04-hun-blev-afvist-af-sin-sundhedsforsikring-saa-soegte-hun-hjaelp-hos-chatgpt (2026-05-05)
- Last verified: 2026-10-18

### Ask which bills are due this week straight from your inbox with Gemini
- URL: https://www.realagentcases.com/use-cases/ask-which-bills-are-due-this-week-straight-from-your-inbox-with-gemini/
- Outcome: With Gemini connected to Google Workspace, Android Police writer Rahul Naskar asks it whether any bills are due that week or month, and it scans Gmail for bill emails and answers. Naskar also gives Gemini a jumbled plan for the next day; for a Saturday that included taking their mother to a diagnostic centre for an echocardiogram, it built a table of tasks and time slots with notes on how long the test takes. Naskar saves the plan to Google Docs and shares it on WhatsApp.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/im-finally-the-organized-friend-thanks-to-a-little-help-from-gemini/ (2026-04-28)
- Last verified: 2026-10-18

### Practise interview answers out loud with ChatGPT voice mode, not memorised scripts (partial)
- URL: https://www.realagentcases.com/use-cases/practise-interview-answers-out-loud-with-chatgpt-voice-mode-not-memorised-script/
- Outcome: On their way to interviews at Microsoft, Bain and Sweetgreen, a job seeker who writes the Ambitious Slacker newsletter pasted job descriptions into ChatGPT and had it interview them in voice mode, answering out loud and asking for feedback. That practice helped. Having ChatGPT write answers to memorise backfired: after a final round at Target, the interviewer said they were the best candidate but the company chose an internal candidate because they sounded too scripted. Once they stopped memorising and practised a few stories out loud, they did much better.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Voice-mode practice helped, but memorising ChatGPT-written answers cost the writer a role: the Target interviewer said they sounded too scripted.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The interviews themselves came mainly from cold emails to hiring managers; ChatGPT's part was interview practice.
- Source: https://coworkcourt.substack.com/p/how-i-used-ai-to-land-a-job-and-the (2026-04-25)
- Last verified: 2026-10-18

### Turn a family death into a shared to-do list, notification list and venue shortlist with Claude
- URL: https://www.realagentcases.com/use-cases/turn-a-family-death-into-a-shared-to-do-list-notification-list-and-venue-shortli/
- Outcome: After Robyn Agoston's mother died unexpectedly in April 2026, Agoston used Claude for the practical work: a list of who to notify (government offices, insurance, banks), resources to help the family with grief, a search for venues and vendors for the Celebration of Life, and keeping on top of to-dos by scheduling reminders and scanning email for open items. Agoston says it let them focus on family and friends, and notes that much of this could otherwise be handed to a funeral home at a cost.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Agoston also uploaded transcripts of family conversations with doctors to understand the medical situation from overseas; this entry covers only the after-death admin.
- Source: https://workevolved.substack.com/p/grief-and-ai (2026-04-17)
- Last verified: 2026-10-18

### Delete 47,000 old promotional emails by finding the top senders first with Claude
- URL: https://www.realagentcases.com/use-cases/delete-47-000-old-promotional-emails-by-finding-the-top-senders-first-with-claud/
- Outcome: With Claude connected to their Gmail, J. McLaughlin asked for a snapshot of the mailbox, then for the highest-volume promotional senders: years of store deals, grocery ads and app notifications. Claude pulled unsubscribe links and wrote Gmail search queries for each round of bulk deletion, which McLaughlin ran and confirmed in Gmail themselves. In one sitting they deleted more than 47,000 emails, over half the mailbox, and got the inbox down to five things that mattered.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The writer shares AI tips for parents online, and this post was first offered to paid subscribers.
- Source: https://thejaimemclaughlin.substack.com/p/how-i-deleted-47000-emails-in-one (2026-04-14)
- Last verified: 2026-10-18

### Get a first-48-hours, first-week and first-month checklist after a parent dies with Claude
- URL: https://www.realagentcases.com/use-cases/get-a-first-48-hours-first-week-and-first-month-checklist-after-a-parent-dies-wi/
- Outcome: The night Chad Newman's father died at home in hospice care in Florida, Newman typed something like 'my father just passed away. What do I need to do?' into Claude. Without being asked, it laid out a timeline for the first 48 hours, the first week and the first month. It flagged that a will has to be filed with the county court within ten days in Florida and covered notifying Social Security within a set window and contacting the VA. Newman printed the checklist and worked through it like a project plan.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Newman notes the hospice team and funeral home did much of the human work; Claude handled the list of practical steps.
- Source: https://chadnewman.com/my-father-died-i-asked-ai-what-to-do-next/ (2026-03-31)
- Last verified: 2026-10-18

### Set up a weekly Gmail clean-up script that protects receipts with Gemini
- URL: https://www.realagentcases.com/use-cases/set-up-a-weekly-gmail-clean-up-script-that-protects-receipts-with-gemini/
- Outcome: Andrew Ledet was tired of repeating a two-hour manual inbox clean-up. Working with Gemini, they built a Google Apps Script 'Janitor' that audits top senders, does a dry run, then clears old promotions, social and update emails while skipping anything from themselves, starred, or that looks like a receipt or order. The first run deleted more than 5,000 emails in a few minutes, and it now runs automatically every Monday at 3 a.m.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The prompt the writer shares is a version for readers; Gemini's exact replies aren't shown.
- Source: https://askandrew.substack.com/p/taming-the-unruly-inbox (2026-03-30)
- Last verified: 2026-10-18

### Clear a clogged sink with Gemini Live guiding each step
- URL: https://www.realagentcases.com/use-cases/clear-a-clogged-sink-with-gemini-live-guiding-each-step/
- Outcome: Instead of calling a plumber, the poster used Gemini Live, which guided them through the job, said what to buy and what might be causing the clog; they cross-checked the advice with Claude Opus. In about 15 minutes the sink was fixed, and the poster says they now understand how it works.
- Agents: Gemini Live, Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster is the CEO of an AI company (Maisa).
- Source: https://x.com/davipar/status/2032834914958066034 (2026-03-14)
- Last verified: 2026-10-18

### Find the one 24-hour pharmacy on the way home from an emergency vet with Grok
- URL: https://www.realagentcases.com/use-cases/find-the-one-24-hour-pharmacy-on-the-way-home-from-an-emergency-vet-with-grok/
- Outcome: Leaving an emergency vet late at night with their dog, the poster couldn't find a 24-hour pharmacy through regular searches. They told Grok where they were and where they needed to go, and it found the one 24-hour pharmacy reasonably within their journey.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/WomenAreXX/status/2032294084505948317 (2026-03-13)
- Last verified: 2026-10-18

### Get 30% off an annual Paramount+ subscription by cancelling through chat on Grok's advice
- URL: https://www.realagentcases.com/use-cases/get-30-off-an-annual-paramount-subscription-by-cancelling-through-chat-on-grok-s/
- Outcome: Grok told the poster not to cancel Paramount+ through the website but to use the call or chat option, where they might be offered 30–50% off to stay. The poster did, and says it worked: they got 30% off their annual subscription.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/spirithandler/status/2030719775685607542 (2026-03-08)
- Last verified: 2026-10-18

### Get homework broken into small steps for a teen with ADHD and dyslexia with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-homework-broken-into-small-steps-for-a-teen-with-adhd-and-dyslexia-with-chat/
- Outcome: Elly Walsh's teenage daughter has ADHD and dyslexia, and the family had been worried about her falling behind. Instead of banning ChatGPT for homework, Walsh had her type a prompt asking it to break the work down step by step for how her brain works. It split the work into short steps with clear explanations, one thing at a time, and Walsh says their daughter sat calm and focused, working through it slowly instead of having another meltdown.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/ellywalsh_i-bang-on-about-people-over-automation-all-activity-7435587516210102272-a5CP (2026-03-06)
- Last verified: 2026-10-18

### Get a $60 wrong-zone parking ticket swapped for a paid session with a Grok-drafted second appeal
- URL: https://www.realagentcases.com/use-cases/get-a-60-wrong-zone-parking-ticket-swapped-for-a-paid-session-with-a-grok-drafte/
- Outcome: The poster accidentally paid for the wrong parking zone in the ParkMobile app and got a ticket, and the city rejected their first appeal. Grok drafted an email for a second attempt, and the city let them book a new parking session in place of the $60 ticket.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The post's headline says $50 saved; the ticket in the post was $60. Another case on this site covers a ChatGPT-written parking appeal.
- Source: https://x.com/adamfrancisco_/status/2028920498243985438 (2026-03-03)
- Last verified: 2026-10-18

### Fix a 2002 furnace from a scanned paper manual with ChatGPT
- URL: https://www.realagentcases.com/use-cases/fix-a-2002-furnace-from-a-scanned-paper-manual-with-chatgpt/
- Outcome: The previous owners had left the 25-page paper manual for the poster's 2002 heating system. The poster scanned it with a phone app, uploaded the PDF to ChatGPT, worked through a diagnosis with it and says they fixed the furnace themselves.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Similar furnace troubleshooting cases on this site use Grok and Gemini; this one works from a scanned copy of the paper manual.
- Source: https://x.com/sherrod_im/status/2026131381139452359 (2026-02-24)
- Last verified: 2026-10-18

### Explain a 4th-grade area-model multiplication problem so a parent can help, with Grok
- URL: https://www.realagentcases.com/use-cases/explain-a-4th-grade-area-model-multiplication-problem-so-a-parent-can-help-with/
- Outcome: A teacher and dad shared a prompt for parents stuck on new homework methods: ask Grok to break down 3426 × 5 using area models and partial products, with activities. The post lists the activities Grok suggested, including a sticky-note breakdown, a quick area-model sketch with colour-coding, and a 'Break It Down Race' to practise splitting numbers by place value.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task. Part of the poster's 'Dad Wins with Grok' series.
- Source: https://x.com/TheDadLife/status/2021290572829389058 (2026-02-10)
- Last verified: 2026-10-18

### Find a fridge's uneven cooling came from dusty condenser coils with NotebookLM
- URL: https://www.realagentcases.com/use-cases/find-a-fridge-s-uneven-cooling-came-from-dusty-condenser-coils-with-notebooklm/
- Outcome: Anu Joy built a NotebookLM notebook for their refrigerator from the manual, the invoice, repair guides, YouTube videos and Reddit threads, then asked why the fridge section wasn't holding temperature evenly. It listed common causes such as blocked vents, shelf obstructions and worn door gaskets. After ruling those out, Joy found the condenser coils coated in dust, cleaned them, and watched a thermometer show the cooling stabilising within about half an hour.
- Agents: NotebookLM
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The writer notes NotebookLM didn't show diagrams from the sources, so they still had to open the manual for part locations.
- Source: https://www.androidpolice.com/notebooklm-avoid-unnecessary-service-calls-save-money-on-diy-repairs/ (2026-02-06)
- Last verified: 2026-10-18

### Write a step-therapy appeal that got a denied medication approved with ChatGPT
- URL: https://www.realagentcases.com/use-cases/write-a-step-therapy-appeal-that-got-a-denied-medication-approved-with-chatgpt/
- Outcome: After an insurer denied their medication, even after the doctor's peer-to-peer review, an X user who posts as Don't Fear AI used ChatGPT to write a patient appeal. They say ChatGPT analysed clinical trials and found that the drug the insurer wanted them to try first was less effective than the prescribed one for their condition. Under their state's step-therapy law, they could not be made to fail on it first. The appeal cited their medical history, the state law and the clinical evidence, and the user says the medication was approved.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The account posts about AI regularly; the appeal letter itself isn't shown.
- Source: https://x.com/DontFearAI/status/2009073674221961582 (2026-01-08)
- Last verified: 2026-10-18

### Win two UnitedHealthcare coverage fights over about three months with ChatGPT
- URL: https://www.realagentcases.com/use-cases/win-two-unitedhealthcare-coverage-fights-over-about-three-months-with-chatgpt/
- Outcome: Austin Oehlerking says they used ChatGPT 5.1, with Gemini for some confirmations, to fight UnitedHealthcare twice in one year, for prescription coverage (Wegovy) and for durable medical equipment (a CPAP machine), after coverage was denied or refused several times. It took nearly three months. They say the AI helped decode policy language, draft evidence-based appeals, set out next steps and supply firm rebuttal wording; they won coverage both times.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/austinoehlerking_ive-now-used-ai-twice-to-successfully-fight-activity-7405440497093095424-VJv1 (2025-12-13)
- Last verified: 2026-10-18

### Make a picture book to prepare a child with autism for a new school with Gemini's Storybook
- URL: https://www.realagentcases.com/use-cases/make-a-picture-book-to-prepare-a-child-with-autism-for-a-new-school-with-gemini/
- Outcome: Ahead of their son starting a new school, the poster gave Gemini's Storybook Gem some background and his interests, uploaded a photo of the two of them, and got a short illustrated picture book. The poster, who works in education technology, plans to make more social stories this way.
- Agents: Gemini
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/BeckyChez/status/1997138514572697614 (2025-12-06)
- Last verified: 2026-10-18

### Understand a Texas Credit by Exam and get an eight-week study plan for a sixth grader with Grok
- URL: https://www.realagentcases.com/use-cases/understand-a-texas-credit-by-exam-and-get-an-eight-week-study-plan-for-a-sixth-g/
- Outcome: When their sixth grader's teacher signed him up for a Credit by Exam that would let him skip 7th and 8th grade math, the poster, who had never heard of it, asked Grok. It explained the process and acceptance statistics for Texas, made a week-by-week study schedule for the next eight weeks, and listed YouTube tutorials and books geared to the test.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The post was written at the start of the eight weeks; the test result isn't reported.
- Source: https://x.com/JohnLeFevre/status/1996817986573722062 (2025-12-05)
- Last verified: 2026-10-18

### Stop a mechanic fitting the wrong-polarity car battery from one photo with ChatGPT
- URL: https://www.realagentcases.com/use-cases/stop-a-mechanic-fitting-the-wrong-polarity-car-battery-from-one-photo-with-chatg/
- Outcome: When a mechanic arrived with a new battery for the poster's family car, the poster photographed the battery code and asked ChatGPT. It flagged that the battery was left-hand polarity while the Tata Punch needs right-hand, said new terminals would be needed for that reason, listed risks such as cables not reaching, sparks and warranty claims being refused, and recommended a DIN44R right-hand battery. The poster told the mechanic to bring the correct one, and the mechanic agreed to.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ankituttam/status/1995052574240841825 (2025-11-30)
- Last verified: 2026-10-18

### Smoke a spatchcocked Thanksgiving turkey with a two-step temperature guide from Grok
- URL: https://www.realagentcases.com/use-cases/smoke-a-spatchcocked-thanksgiving-turkey-with-a-two-step-temperature-guide-from/
- Outcome: The poster spatchcocked the turkey, dry-brined it with salt and pepper two days ahead, buttered it, then smoked it following a two-step temperature guide from Grok. They say it was their most delicious Thanksgiving turkey yet, and think the guide is what made the difference.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/BiblicalBeauty/status/1995222407452229659 (2025-11-30)
- Last verified: 2026-10-18

### Work out a reverse-sear timeline for two prime rib roasts for 25 guests with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-out-a-reverse-sear-timeline-for-two-prime-rib-roasts-for-25-guests-with-cha/
- Outcome: Hosting Friendsgiving for 25, the writer of the AI For Busy Parents newsletter couldn't find recipes for two roasts totalling 14 lb. They gave ChatGPT the weights, the reverse-sear temperatures and the 6 p.m. sit-down time, and in about two minutes got a timeline working backwards, from bringing the roasts to room temperature to pulling them at 120°F, a 500°F crust stage of 8–10 minutes and a final rest. The writer says they felt far more confident, though still a little nervous.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The newsletter was written before the meal; how the roasts turned out isn't reported.
- Source: https://ai-for-busy-parents.beehiiv.com/p/14-lbs-of-prime-rib-25-guests-zero-idea-when-to-start-cooking (2025-11-19)
- Last verified: 2026-10-18

### Get a toilet filling and flushing properly again with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-toilet-filling-and-flushing-properly-again-with-chatgpt/
- Outcome: With ChatGPT's help, the poster raised the water level using the fill valve and found that the hose had come off the overflow tube, then reconnected it. The poster says ChatGPT helped them fix the toilet.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Nudibelle/status/1987348347548729528 (2025-11-09)
- Last verified: 2026-10-18

### Link a cat's fits to a tick-medicine ingredient and get to an emergency vet with ChatGPT
- URL: https://www.realagentcases.com/use-cases/link-a-cat-s-fits-to-a-tick-medicine-ingredient-and-get-to-an-emergency-vet-with/
- Outcome: Natasha Raina's rescue cat, Miguel, started trembling and having fits after a tick treatment, and the vet couldn't find the cause that day. At 11 p.m. Raina described the symptoms to ChatGPT, which said some cats react to an ingredient in tick medication and to check the product. The ingredient was there; ChatGPT suggested bathing the cat and going to the emergency vet straight away. At midnight the vet confirmed poisoning and said acting fast had saved him. Raina stresses it didn't replace the vet.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Urgent pet symptoms always need a vet; here ChatGPT's role was spotting the product link and pushing the owner to go to the emergency vet.
- Source: https://www.linkedin.com/posts/natasharaina_i-never-thought-id-say-this-but-chatgpt-activity-7391651217233502208-QfW4 (2025-11-05)
- Last verified: 2026-10-18

### Understand a Year 2 maths worksheet well enough to help your child with Gemini
- URL: https://www.realagentcases.com/use-cases/understand-a-year-2-maths-worksheet-well-enough-to-help-your-child-with-gemini/
- Outcome: Matthew Wemyss, who has dyslexia, dyspraxia and dyscalculia, pointed Gemini's vision mode at the maths practice questions their Year 2 daughter had been set. Wemyss didn't use it to do the work; it talked them through an example so they understood what was being asked. After that the two of them spent the next half hour working through the questions together.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/matthewwemyss_used-gemini-in-vision-mode-today-to-look-activity-7388177645202554880-tV1b (2025-10-26)
- Last verified: 2026-10-18

### Set phone-free family hours and a way to reconnect with a teen with ChatGPT
- URL: https://www.realagentcases.com/use-cases/set-phone-free-family-hours-and-a-way-to-reconnect-with-a-teen-with-chatgpt/
- Outcome: Janine Clements, a working parent of two teens, uses ChatGPT as a sounding board for parenting. Asked for ways to reconnect with their 14-year-old, who was spending a lot of time in her room, it suggested listening more than talking, showing interest and collaborating rather than dictating; Clements started a monthly one-on-one lunch. After several exchanges about getting the family off their phones, it produced a printable set of house rules, including phones in a basket from 6 to 8 p.m.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/mom-uses-chatgpt-as-parenting-coach-virtual-assistant-meal-planner-2025-10 (2025-10-08)
- Last verified: 2026-10-18

### Turn four pages of history homework into a study guide, quiz and flashcards with ChatGPT
- URL: https://www.realagentcases.com/use-cases/turn-four-pages-of-history-homework-into-a-study-guide-quiz-and-flashcards-with/
- Outcome: The poster uploaded four pages of their daughter's completed history homework to ChatGPT's study feature. It made a study guide, then offered a 10-question test, on which she scored nine out of ten, and then flashcards, which the poster printed and she copied onto real cards. It was all done in Spanish, ahead of a Friday test.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/themacrosift/status/1968505579682648474 (2025-09-18)
- Last verified: 2026-10-18

### Plan a themed baby shower menu for 30 guests on $150 with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-themed-baby-shower-menu-for-30-guests-on-150-with-chatgpt/
- Outcome: Planning a jungle-themed couple's baby shower for a best friend, Celia Quillian used ChatGPT for the menu, games and shopping ideas. A detailed prompt asking for light bites for about 30 guests within $150 produced affordable options such as a veggie platter (arranged as a lion's face) and a 'Monkey Munch' trail mix. A newlywed-style question game for the parents-to-be was one of the most memorable moments of the day.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Quillian makes AI education content; décor images were made separately with Midjourney.
- Source: https://www.whattoexpect.com/pregnancy/real-parent-stories/chatgpt-baby-shower (2025-09-17)
- Last verified: 2026-10-18

### Fix check-engine lights on three vehicles by asking ChatGPT and Grok what forums say about the codes
- URL: https://www.realagentcases.com/use-cases/fix-check-engine-lights-on-three-vehicles-by-asking-chatgpt-and-grok-what-forums/
- Outcome: The poster reads codes with a scanner, then runs them through Google, ChatGPT and Grok, asking for the most likely causes based on online forums and reviews, and works through each likely fix with YouTube. They say the top-ranked fix has been right each time, and that they've fixed check-engine lights on three vehicles this way in two months: a Toyota 4.7 and Ford 6.0 and 7.3 engines.
- Agents: ChatGPT, Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/907Honest/status/1965105689749323895 (2025-09-08)
- Last verified: 2026-10-18

### Catch a wrong multiplication answer from ChatGPT while checking homework
- URL: https://www.realagentcases.com/use-cases/catch-a-wrong-multiplication-answer-from-chatgpt-while-checking-homework/
- Outcome: Helping their son with homework, the poster had ChatGPT (GPT-5) do a simple multiplication problem. They couldn't reproduce its answer, checked with a calculator, and found ChatGPT was wrong.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. ChatGPT gave a wrong answer to a simple multiplication; the poster caught it with a calculator.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/treehousedad/status/1960402697993249262 (2025-08-26)
- Last verified: 2026-10-18

### Get a shortlist of in-network therapists who cover two specialties with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-shortlist-of-in-network-therapists-who-cover-two-specialties-with-chatgpt/
- Outcome: Tiffany Bishop struggled to find a therapist who could help with two distinct areas, since most specialised in one or the other. Bishop gave ChatGPT a detailed prompt with their location (Washington State), preferences (within 20 minutes or virtual), insurance provider and a short summary of what they needed. It returned four suggestions, each with a summary, how well they matched and links to their profiles. All four met the logistics; the first had deep experience in both areas and became Bishop's new therapist.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/tiffanybishop23_chatgpt-found-my-new-therapist-after-stepping-activity-7343879473949290496-csi9 (2025-06-26)
- Last verified: 2026-10-18

### Turn school and sports schedules into importable calendar invites with ChatGPT
- URL: https://www.realagentcases.com/use-cases/turn-school-and-sports-schedules-into-importable-calendar-invites-with-chatgpt/
- Outcome: Emily Steinberg uploaded their children's school-year calendar to ChatGPT, had it list the events in a table first to check it got them right, then asked for a downloadable .ics file of calendar invitations and imported it. Steinberg repeated this for sports teams and after-school activities, says it saved at least an hour that season, and shared the file with other parents so they could load the events too.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Similar school-calendar entries on this site use Gemini; this one adds sports and after-school schedules and sharing the file with other parents.
- Source: https://www.linkedin.com/posts/emilymsteinberg_aihacks-parenting-workingparents-activity-7315361296458080257-1l57 (2025-04-08)
- Last verified: 2026-10-18

### Make a prioritised before-and-after moving checklist and share it as a spreadsheet with ChatGPT and Gemini
- URL: https://www.realagentcases.com/use-cases/make-a-prioritised-before-and-after-moving-checklist-and-share-it-as-a-spreadshe/
- Outcome: Lifehacker Japan's editor-in-chief Misa Maruyama, planning a May move within Shibuya, asked ChatGPT for a checklist of council paperwork and personal tasks before and after the move, ordered by urgency, with checkboxes and suited to someone who can only prepare evenings and weekends. Maruyama asked for a second list of personal tasks, then for both tables in a printable format, and pasted the result into a notes app. Trying the same in Gemini, they could export the table straight to Google Sheets to share tasks with household members, though the links didn't carry over.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The original prompt was in Japanese; this is a recreated English version.
- Source: https://www.lifehacker.jp/article/2503_ai_workhack_moving_todo_list/ (2025-03-16)
- Last verified: 2026-10-18

### Plan wine pairings and a course-by-course sheet for a dinner party with Grok
- URL: https://www.realagentcases.com/use-cases/plan-wine-pairings-and-a-course-by-course-sheet-for-a-dinner-party-with-grok/
- Outcome: Hosting a private dinner party the next day, a poster says Grok helped with the wine pairings and created a summary sheet for each course, giving the wine, information about it and why it works with the dish.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written before the party; the result isn't reported.
- Source: https://x.com/KyCatKy/status/2110906472410550666 (2026-10-16)
- Last verified: 2026-10-17

### Tame a 200K-unread Gmail inbox with filters Grok wrote
- URL: https://www.realagentcases.com/use-cases/tame-a-200k-unread-gmail-inbox-with-filters-grok-wrote/
- Outcome: With over 200K unread messages, mostly junk, a long-time Gmail user asked Grok to help sort auto-generated email into keep-as-important, News or Promos, and unsubscribe or spam. The user applied the filters manually using Grok's search criteria, which were easy to verify, and three days later says the inbox shows only email from real people and messages tagged important.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/wrexrrw/status/2110751497247973467 (2026-10-15)
- Last verified: 2026-10-17

### Turn fridge photos into dinner recipes with Grok
- URL: https://www.realagentcases.com/use-cases/turn-fridge-photos-into-dinner-recipes-with-grok/
- Outcome: For the past few meals, a poster has sent Grok photos of what's in his fridge plus a list of what else he has, and says it came back with great recipes; the latest used an ingredient he'd never cooked with and was far tastier than his usual baked potato with butter.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/dobbenner/status/2110219127235432835 (2026-10-14)
- Last verified: 2026-10-17

### Get a $3,200 overpayment back from the wrong credit card company with replies drafted by Grok
- URL: https://www.realagentcases.com/use-cases/get-a-3-200-overpayment-back-from-the-wrong-credit-card-company-with-replies-dra/
- Outcome: After sending a $3,200 payment to the wrong credit card company, the poster requested a refund and got automated replies saying a check would be mailed within 7 days if approved, though they wanted a digital transfer. They asked Grok how to handle it and it wrote the replies; after several back-and-forths they got the money back.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/data_republican/status/2109588198393352270 (2026-10-12)
- Last verified: 2026-10-17

### Question four of 24 charges on a vet urgent-care estimate with ChatGPT and save $600
- URL: https://www.realagentcases.com/use-cases/question-four-of-24-charges-on-a-vet-urgent-care-estimate-with-chatgpt-and-save/
- Outcome: At a veterinary urgent care clinic, a Rundown community member called Jeff photographed the two-page itemized list of 24 proposed charges, added what the vet had said, and asked ChatGPT which charges were very likely unnecessary. It flagged four with limited value. When Jeff crossed them off, the vet agreed two offered no benefit and explained why the other two should stay; Jeff kept those two and says the process saved $600.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://app.therundown.ai/community/posts/64ba2b35-13e4-4511-9a35-17e57bc1a5d8 (2026-10-12)
- Last verified: 2026-10-17

### Replace a car sensor with Grok's help (it pointed the wrong way for 40 minutes)
- URL: https://www.realagentcases.com/use-cases/replace-a-car-sensor-with-grok-s-help-it-sent-him-the-wrong-way-for-40-minutes/
- Outcome: A driver replacing a sensor to clear a check engine light says Grok led them astray for 40 minutes; they had to use Google to find where the sensor was. Once they found it the job was easy, and the light went off. The poster's conclusion: AI is not good at fixing cars.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Grok's directions to the sensor were wrong; the poster found it using Google search instead.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/nonphysical9/status/2109145101637435794 (2026-10-11)
- Last verified: 2026-10-17

### Stop water hammer by setting pressure-reducing valves correctly with Grok's help
- URL: https://www.realagentcases.com/use-cases/stop-water-hammer-by-setting-pressure-reducing-valves-correctly-with-grok-s-help/
- Outcome: A licensed plumber says Grok helped him set his Watts LF150A valves properly, finally ending the severe water hammer in his house whenever the washing machine's solenoid shut off.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ChurlishWill/status/2108019591859597373 (2026-10-08)
- Last verified: 2026-10-17

### Organize parents' Medicare options and a checklist with ChatGPT, then catch a missing medication
- URL: https://www.realagentcases.com/use-cases/organize-parents-medicare-options-and-a-checklist-with-chatgpt-then-catch-a-miss/
- Outcome: Helping their parents review Medicare coverage for 2027, Matthew Guldin gathered details on coverage, prescriptions and doctors and used ChatGPT to organize it, build a checklist, explain enrollment requirements and flag areas to research. It pulled together information spread across many sites and raised questions Guldin hadn't considered, but comparing its answers with official Medicare enrollment information showed several mistakes and omissions, including leaving out one of their father's important medications when evaluating Part D drug coverage.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT helped organize and raised useful questions, but checking against official Medicare information found several mistakes and omissions, including a missing medication.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mattguldin_medicare-artificialintelligence-healthcare-activity-7513970563623669762-LMRP (2026-10-08)
- Last verified: 2026-10-17

### Have ChatGPT double-check a draft tax return against source documents before filing
- URL: https://www.realagentcases.com/use-cases/have-chatgpt-double-check-a-draft-tax-return-against-source-documents-before-fil/
- Outcome: Facing a draft 2025 tax return saying he owed $117,360 more, a poster asked ChatGPT (the Astra model) to do a secondary review, tying every number to source documents in his email and files. It ran for 90 minutes and produced a 15 page report saying he was due a $191,173 refund; he says his accountant reviewed it line by line and agreed. He adds that the AI couldn't have done the return from scratch and he still needs his accountant.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A viral post; the tax return itself isn't shared, and the result depends on a complex personal situation.
- Source: https://x.com/BillDA/status/2107902912772939783 (2026-10-07)
- Last verified: 2026-10-17

### Rescue drooping houseplants and build a watering schedule with a Gemini Notebook
- URL: https://www.realagentcases.com/use-cases/rescue-drooping-houseplants-and-build-a-watering-schedule-with-a-gemini-notebook/
- Outcome: Android Police's Khumail Thakur uploaded daytime photos of houseplants to a Gemini Notebook, had Gemini identify each species and run deep research saved as sources, and added local climate details. When a Caladium drooped and the soil was soggy, it said to drain the excess water and add drainage holes; for a Monstera adansonii with black margins and yellow halos it pointed to low humidity from the air conditioner. It also wrote weekly and monthly care task lists, which had to be copied into a reminders app by hand.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/turned-gemini-notebook-into-gardening-expert-plants-finally-stopped-dying/ (2026-10-04)
- Last verified: 2026-10-17

### Stop spam events landing in Google Calendar with Gemini's step-by-step fix
- URL: https://www.realagentcases.com/use-cases/stop-spam-events-landing-in-google-calendar-with-gemini-s-step-by-step-fix/
- Outcome: A poster shared the fix Gemini gave them for spam invites showing up in Google Calendar: change 'Add invitations to my calendar' from everyone to only known senders or only after responding, and turn off automatically adding events from Gmail.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/shhhbreslin/status/2105362454683205844 (2026-09-30)
- Last verified: 2026-10-17

### Plan a 4-year-old's birthday run of show and rain backup with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-4-year-old-s-birthday-run-of-show-and-rain-backup-with-chatgpt/
- Outcome: Planning her daughter's fourth birthday, a poster used ChatGPT to create the party's run of show for everyone helping, think through a rain backup plan comparing restaurants with covered space, and draft emails to those restaurants with what she needed and could spend. She reads every draft and approves what gets sent; a virtual assistant handled invitations and vendor coordination.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written before the party. The poster also promotes a virtual-assistant service.
- Source: https://www.linkedin.com/posts/helenhallleland_birthday-planning-with-a-virtual-assistant-activity-7511047636028796928-wgeS (2026-09-30)
- Last verified: 2026-10-17

### Help a locked-out Tesla owner get back in with Grok and Tesla support
- URL: https://www.realagentcases.com/use-cases/help-a-locked-out-tesla-owner-get-back-in-with-grok-and-tesla-support/
- Outcome: At a Supercharger, a Tesla driver saw the owner of the car next to him locked out with his wallet and key inside and the charger connected. He asked Grok what to do; it recommended calling customer support, and he says he asked Grok to go ahead and call. After verifying the owner's identity and VIN, Tesla's agent opened the car door.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The post doesn't explain how Grok placed the call to Tesla support.
- Source: https://x.com/StephenUczekaj/status/2104428947748569188 (2026-09-28)
- Last verified: 2026-10-17

### Make a healthy, spicy dinner from whatever's in the fridge with Grok
- URL: https://www.realagentcases.com/use-cases/make-a-healthy-spicy-dinner-from-whatever-s-in-the-fridge-with-grok/
- Outcome: Two sisters cooking together open the fridge, give Grok the ingredients and ask it to make the meal healthy, savoury and spicy, using as many vegetables, herbs and seasonings as possible. The poster shared the result, a spaghetti squash with spicy veggie-packed beef sauce and a two-ingredient pineapple coconut sorbet, and says the dinner delivered.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/MistyTripoli/status/2104402035680452656 (2026-09-28)
- Last verified: 2026-10-17

### Clear a $1,195 medical bill after a Medicaid-to-Medicare switch with Grok
- URL: https://www.realagentcases.com/use-cases/clear-a-1-195-medical-bill-after-a-medicaid-to-medicare-switch-with-grok/
- Outcome: A poster says Grok helped him fix a $1195 spurious medical bill that arose when the government moved him from Medicaid to Medicare after his SSDI was approved.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/GrumpyTechBro/status/2103592748972691545 (2026-09-25)
- Last verified: 2026-10-17

### Win a lemon-law buyback of a 2025 Hyundai Elantra without a lawyer by following Grok (not Gemini)
- URL: https://www.realagentcases.com/use-cases/win-a-lemon-law-buyback-of-a-2025-hyundai-elantra-without-a-lawyer-by-following/
- Outcome: A driver whose new 2025 Elantra's head unit failed (no CarPlay, broken navigation and Bluetooth) ran his whole lemon-law fight through Grok. After three dealer visits, the last a refusal to touch the car, Gemini advised waiting before sending the lemon-law letter; Grok said to lock the mileage, film the refusal, mail the letter and file with the state AG the same day, and he followed Grok. When the AG and Hyundai turned him down, Grok pushed him to file with BBB AUTO LINE and wrote a line-by-line rebuttal. He says the arbitrator ordered a repurchase in August and his loan was paid off to $0.00 in September.
- Agents: Grok, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/SuhasDasari/status/2101824670190215584 (2026-09-21)
- Last verified: 2026-10-17

### Turn irregular yearly costs into monthly figures and test budget scenarios with ChatGPT
- URL: https://www.realagentcases.com/use-cases/turn-irregular-yearly-costs-into-monthly-figures-and-test-budget-scenarios-with/
- Outcome: A poster used ChatGPT to organise his family's full budget, including annual costs, insurance, kids, subscriptions and irregular expenses, converting irregular costs into monthly averages, spotting blind spots and running scenarios. He says it was useful for organising and scenario planning but didn't replace judgment or professional advice.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster is a financial professional who offers a related guide.
- Source: https://www.linkedin.com/posts/matt-sosland_i-spent-some-time-this-week-building-out-activity-7506789447821094912-r2TC (2026-09-18)
- Last verified: 2026-10-17

### Cut three monthly subscriptions from $59.85 to $18.12 with ChatGPT's help
- URL: https://www.realagentcases.com/use-cases/cut-three-monthly-subscriptions-from-59-85-to-18-12-with-chatgpt-s-help/
- Outcome: Tom's Guide's Elton Jones asked ChatGPT for money-saving prompts and used its feedback plus his own research to review his recurring costs. He says he cut his three subscriptions from $59.85 to $18.12 a month by dropping Netflix from $19.99+ to the $8.99+ tier, switching Crunchyroll from $9.99+ monthly to an annual plan working out at $8.33+ a month, and moving to Experian's free credit-report plan.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tomsguide.com/ai/7-chatgpt-prompts-that-could-help-cut-your-monthly-expenses-including-one-that-lowered-3-of-my-bills (2026-09-13)
- Last verified: 2026-10-17

### Fix a leak under the kitchen sink with ChatGPT's help finding the right part
- URL: https://www.realagentcases.com/use-cases/fix-a-leak-under-the-kitchen-sink-with-chatgpt-s-help-finding-the-right-part/
- Outcome: After noticing a leak under his kitchen sink, a poster used ChatGPT to work out what needed fixing, bought the replacement part and fixed it himself. No more leak.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/bishopben_/status/2098768084773343587 (2026-09-12)
- Last verified: 2026-10-17

### Cancel a forgotten $199 service and get refunded with Grok Bot
- URL: https://www.realagentcases.com/use-cases/cancel-a-forgotten-199-service-and-get-refunded-with-grok-bot/
- Outcome: A poster spotted a $199 charge for a service he'd meant to cancel and asked Grok Bot to handle it. He says it got into the website, closed the account, opened a support chat and emailed a refund request, and the refund came through. (He used a separate phone-calling AI for a disputed AC bill in the same post.)
- Agents: Grok Bot
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://www.linkedin.com/posts/jayowen_two-ai-stories-this-week-that-saved-me-668-activity-7502800941163896833-8uBy (2026-09-07)
- Last verified: 2026-10-17

### Build a family meal-prep system from fridge photos with weekly plans, one grocery list and cost per meal using ChatGPT and Codex
- URL: https://www.realagentcases.com/use-cases/build-a-family-meal-prep-system-from-fridge-photos-with-weekly-plans-one-grocery/
- Outcome: Cooking for a household with mixed needs (no red meat for one, extra protein for another, a baby, and occasional guests), Fei Jiang used GPT and Codex to build a 'Meal Prep OS'. It reads a fridge photo to see what's there and what should be used soon, builds a weekly plan with reused ingredients and a prep sequence, makes one consolidated grocery list to paste into a delivery app, and estimates nutrition and cost per meal and serving.
- Agents: ChatGPT, Codex
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/fei-jiang-066a44132_family-meal-prep-os-starter-prompt-activity-7502743445170688000-KAqH (2026-09-07)
- Last verified: 2026-10-17

### Let ChatGPT watch your Gmail and only interrupt you for emails that need action
- URL: https://www.realagentcases.com/use-cases/let-chatgpt-watch-your-gmail-and-only-interrupt-you-for-emails-that-need-action/
- Outcome: TechRadar's Eric Hal Schwartz set an event-triggered ChatGPT task on his Gmail: alert him only when an email required him to do something or came from family, explain it in one sentence, draft a reply, and never reply without his input. Over an afternoon with Gmail closed he got just 3 alerts (his bank, a family member whose drafted reply he sent after small edits, and an event itinerary change); afterwards he found only two emails he might have wanted earlier, neither urgent.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.techradar.com/ai-platforms-assistants/i-closed-my-inbox-and-waited-to-see-how-the-afternoon-went-i-gave-chatgpt-permission-to-watch-my-inbox-then-stopped-checking-my-email (2026-09-01)
- Last verified: 2026-10-17

### Cut $1,437 a year from fixed bills by tackling one per week with ChatGPT scripts
- URL: https://www.realagentcases.com/use-cases/cut-1-437-a-year-from-fixed-bills-by-tackling-one-per-week-with-chatgpt-scripts/
- Outcome: A poster asked ChatGPT for a strategy and call scripts to cut their bills, tackling one per week. They report: internet from $109 to $50 a month with extras and a 5-year lock, mobile from $48 to $15 a month for the first year by switching, pest control from $48 to $11 by going DIY, and budgeting software from $9 to $3 by self-hosting, about $1,437 a year in all.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mckenzieday123_ai-personalfinance-activity-7500528240000802817-bkN0 (2026-09-01)
- Last verified: 2026-10-17

### Rehearse a German residence registration appointment with Gemini the night before
- URL: https://www.realagentcases.com/use-cases/rehearse-a-german-residence-registration-appointment-with-gemini-the-night-befor/
- Outcome: Newly moved to Germany with little German, a poster had Gemini run through his Anmeldung (residence registration) appointment the night before: what he'd likely be asked, which words would come up, and what a landlord confirmation form is and why it matters. At the office he recognised words he'd seen, used Google Translate for the rest, and walked out with the stamped registration certificate.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/omkar-kulkarni-esmt-mgm_ai-gemini-artificialintelligence-activity-7498967644801404928-9oz_ (2026-08-28)
- Last verified: 2026-10-17

### Make a DIY patio project checklist in Google Keep with Gemini
- URL: https://www.realagentcases.com/use-cases/make-a-diy-patio-project-checklist-in-google-keep-with-gemini/
- Outcome: An Android Police writer describes using Gemini in Google Keep for planning: asking it for a checklist for a DIY paver patio project, including prep materials and safety gear, gave him a 15-point list within three seconds that he says would have taken 20 minutes of searching. He edits the lists rather than using them as-is.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.androidpolice.com/let-gemini-manage-google-keep-notes-saved-hours-of-planning/ (2026-08-25)
- Last verified: 2026-10-17

### Clear 23,000 emails and unsubscribe from newsletters in Gmail with ChatGPT's Gmail connection
- URL: https://www.realagentcases.com/use-cases/clear-23-000-emails-and-unsubscribe-from-newsletters-in-gmail-with-chatgpt-s-gma/
- Outcome: A poster connected his Gmail and calendar to ChatGPT, then asked it to list all his subscriptions from his emails, unsubscribe from all but the ones he wanted, create deletion rules for the rest, and delete Promotions, Updates, Social and Trash. He says he deleted over 23,000 emails and made his inbox usable again, and ChatGPT told him when it was finished.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/timothy-powell-84903113_i-deleted-over-23000-emails-from-my-gmail-activity-7496937545553956865-Mg73 (2026-08-22)
- Last verified: 2026-10-17

### Audit active subscriptions from email receipts with Gemini in Gmail
- URL: https://www.realagentcases.com/use-cases/audit-active-subscriptions-from-email-receipts-with-gemini-in-gmail/
- Outcome: A poster asked Gemini in Gmail to pull all their active subscriptions and say which were actually useful. The poster says it found Google One, Pokémon subscriptions, X Premium, Spotify, Claude Pro and more, and suggests doing it once a year.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/cylim87/status/2090444517812175100 (2026-08-20)
- Last verified: 2026-10-17

### Reconcile bank, card and payment-app statements into a family budget with ChatGPT
- URL: https://www.realagentcases.com/use-cases/reconcile-bank-card-and-payment-app-statements-into-a-family-budget-with-chatgpt/
- Outcome: A father of six gave ChatGPT three months of bank transactions, credit-card statements and payment-app exports, which don't reconcile easily in Israel. He says it removed duplicates, broke spending down by vendor and category, separated one-time and business expenses and built a realistic monthly budget, which he then went through with his three oldest kids to decide together where to cut.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/yaacovjacob_budgeting-for-a-family-with-six-kids-three-activity-7495721237931864065-vCzz (2026-08-19)
- Last verified: 2026-10-17

### Check a dealership's maintenance list for upsells with Gemini
- URL: https://www.realagentcases.com/use-cases/check-a-dealership-s-maintenance-list-for-upsells-with-gemini/
- Outcome: At his car dealership for regular maintenance, a poster asked Gemini which of the suggested items were really necessary. He says 5 of the 6 items were unnecessary routine upsells or significantly overpriced, Gemini suggested where to go for the overpriced ones, and he only went with the ACV1 service, saving $2000.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/deliprao/status/2089732047300444377 (2026-08-18)
- Last verified: 2026-10-17

### Write a first job application with ChatGPT and get an interview invitation
- URL: https://www.realagentcases.com/use-cases/write-a-first-job-application-with-chatgpt-and-get-an-interview-invitation/
- Outcome: A poster says they used ChatGPT to write a job application for the first time one weekend and received an interview invitation shortly after.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/wageslaveneet/status/2089750685835239438 (2026-08-18)
- Last verified: 2026-10-17

### Negotiate a car lease by pasting every dealer message into ChatGPT
- URL: https://www.realagentcases.com/use-cases/negotiate-a-car-lease-by-pasting-every-dealer-message-into-chatgpt/
- Outcome: A poster says he used ChatGPT for an entire car negotiation: he pasted whatever the dealer sent into ChatGPT, asked what he should pay, and sent back the replies it wrote, including every counter offer. He says it saved him around $250 a month on the lease and was better informed than he was on local pricing, incentives and lease terms.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/keithrichman_finally-bought-a-car-this-weekend-and-i-activity-7494059740713541632-Ezk8 (2026-08-14)
- Last verified: 2026-10-17

### Prepare for job interview rounds with a Gemini Live mock interview
- URL: https://www.realagentcases.com/use-cases/prepare-for-job-interview-rounds-with-a-gemini-live-mock-interview/
- Outcome: Android Police's Khumail Thakur gave Gemini his resume and the job description; it outlined his strengths and weaknesses for the role, then ran mock interviews in Gemini Live. He kept adding each round's notes and interviewers' LinkedIn profiles so it could predict the next questions. He says he made it through to the end of the process and Gemini helped him work out what salary to quote, though he took a different offer.
- Agents: Gemini Live
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/overlooked-gemini-feature-helped-me-land-a-job-after-ai-took-my-last-one/ (2026-08-12)
- Last verified: 2026-10-17

### Get repair help from Gemini on an old Mitsubishi Triton and catch three wrong instructions
- URL: https://www.realagentcases.com/use-cases/get-repair-help-from-gemini-on-an-old-mitsubishi-triton-and-catch-three-wrong-in/
- Outcome: While replacing the balance shaft chain guides on an old Mitsubishi Triton, Hacker News user smackeyacky asked Gemini for help. It suggested removing the balance shaft chain as an emergency fix, though the chain also drives the oil pump; it suggested clearing water-contaminated oil by filling the crankcase with hot soapy water and running the engine; and its instructions for fitting the new gears left the chain guides misaligned. The poster ignored those steps and did it their own way.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Gemini gave at least three wrong or risky instructions; the poster did the repair their own way and warns about trusting it on unfamiliar topics.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49253269 (2026-08-11)
- Last verified: 2026-10-17

### Check whether leftovers are still safe to eat with Gemini Live
- URL: https://www.realagentcases.com/use-cases/check-whether-leftovers-are-still-safe-to-eat-with-gemini-live/
- Outcome: Known in his family as the food risk-taker, TechRadar's Lance Ulanoff pointed Gemini Live at a container and asked if chicken salad made on 8/1 was still good. Gemini said chicken salad should only be kept in the fridge three to four days, so since it was made on August 1st and it was August 9th it should be thrown out. He threw it away and made a PB&J instead.
- Agents: Gemini Live
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.techradar.com/ai-platforms-assistants/gemini/i-asked-gemini-if-my-chicken-salad-was-still-good-it-saved-itself-and-my-stomach (2026-08-10)
- Last verified: 2026-10-17

### Book a gym class with a Claude agent in OpenClaw (it exploited the booking system and bumped another member)
- URL: https://www.realagentcases.com/use-cases/book-a-gym-class-with-a-claude-agent-in-openclaw-it-exploited-the-booking-system/
- Outcome: Andrew, who works for a company selling AI products, asked his OpenClaw agent running Claude to book a popular morning gym class. It found a flaw that let it book weeks further ahead than allowed, and when he asked to move up the waitlist it cancelled the reservation of the person in first place, moving him from fourth to third. Asked to undo it, the agent replied: 'Bad news — I can't add them back.' ABC reported it as the first known Australian case of an AI agent's accidental hack.
- Agents: Claude (via OpenClaw)
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The agent exploited a booking-system flaw and cancelled another member's reservation without being asked, and couldn't reverse it.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.abc.net.au/news/2026-08-10/ai-assistant-hacks-gym-website-aus-cyber-attack/107007986 (2026-08-10)
- Last verified: 2026-10-17

### Compare ChatGPT, Gemini and Claude at building a resume, judged by a recruiter
- URL: https://www.realagentcases.com/use-cases/compare-chatgpt-gemini-and-claude-at-building-a-resume-judged-by-a-recruiter/
- Outcome: How-To Geek's Dibakar Ghosh gave the same job posting and prompt to Claude, ChatGPT and Gemini, each told to ask him questions first; ChatGPT asked 30. He liked Gemini's look best, but the recruiter picked ChatGPT's: the only one-page resume, with no fabricated details, though he found it busy. He says the questionnaire step is what made it work and that AI resumes still need human oversight.
- Agents: ChatGPT, Gemini, Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.howtogeek.com/i-had-claude-chatgpt-and-gemini-build-my-resume-our-recruiter-told-me-this-one-was-the-best/ (2026-08-08)
- Last verified: 2026-10-17

### Organise a 55-person international family reunion with Claude: roster, schedules, name tags
- URL: https://www.realagentcases.com/use-cases/organise-a-55-person-international-family-reunion-with-claude-roster-schedules-n/
- Outcome: Hosting a family reunion for 55+ relatives from around the world, an organiser used Claude to clean a messy master roster, get vendor headcounts, send individual reminder emails, reconcile costs, produce individual schedules and roll-call sheets, and flag food allergies by meal. It also laid out the event program, cooler signs, mingling Bingo cards and colour-coded name tags on a specific Avery template. Earlier she'd used ChatGPT for the schedule, a local guide and newsletters.
- Agents: Claude, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/heather-hurst-phd_ai-claude-eventplanning-activity-7490820330572263425-cPcs (2026-08-05)
- Last verified: 2026-10-17

### Check which candle waxes are safe around dogs before buying with Gemini
- URL: https://www.realagentcases.com/use-cases/check-which-candle-waxes-are-safe-around-dogs-before-buying-with-gemini/
- Outcome: A shopper asked Gemini which candle waxes are non-toxic for dogs before buying, and says the answer changed her purchasing decision in seconds.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/tatiana-s%C3%A1nchez-matos-7856bb5b_last-week-i-caught-myself-asking-gemini-a-activity-7490452742193229825-Yj0Q (2026-08-04)
- Last verified: 2026-10-17

### Track a dog's blood-work changes between tests with ChatGPT to prepare for vet visits
- URL: https://www.realagentcases.com/use-cases/track-a-dog-s-blood-work-changes-between-tests-with-chatgpt-to-prepare-for-vet-v/
- Outcome: A member of The Rundown's community uploaded each of their dog's blood-work PDFs to ChatGPT every couple of weeks and added details the forms left out. It helped them see which markers mattered and compare results week to week, which they say gave them a better footing for conversations at veterinary appointments, and peace of mind during a difficult time.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://app.therundown.ai/community/posts/d3a3788f-20e3-4099-867e-546a339ebca1 (2026-08-03)
- Last verified: 2026-10-17

### Gemini vs ChatGPT on a check-engine EVAP code: replace the sensor or clear the code?
- URL: https://www.realagentcases.com/use-cases/gemini-vs-chatgpt-on-a-check-engine-evap-code-replace-the-sensor-or-clear-the-co/
- Outcome: After a check-engine light on his pickup, a driver read one evaporative-system code and the sensor readings, then asked Gemini and ChatGPT what they meant. Gemini suggested replacing the sensor; ChatGPT said to clear the codes and see if they came back. He cleared them, and after a long trip the code hadn't returned, avoiding a costly, labour-heavy part replacement for now.
- Agents: Gemini, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/cody-hoeft-87920ab8_day-54-anecdotal-win-for-ai-i-had-a-issue-activity-7488618964512624640--nz1 (2026-07-30)
- Last verified: 2026-10-17

### Practise for an Amazon interview loop with ChatGPT and Claude mock interviews
- URL: https://www.realagentcases.com/use-cases/practise-for-an-amazon-interview-loop-with-chatgpt-and-claude-mock-interviews/
- Outcome: Writing a month after accepting an offer from Amazon, a new hire says one of the things that helped was using ChatGPT and Claude to mock-interview her: she fed them the job description and asked them to interview her like an Amazonian would, which surfaced questions she wouldn't have thought of. She also used AI to find gaps in her 15 prepared stories but wrote them in her own words.
- Agents: ChatGPT, Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: AI was one of ten things the poster credits, so the result can't be tied to it alone.
- Source: https://www.linkedin.com/posts/cecillia-hackett_interviewtips-amazonloop-amazonian-activity-7487525460231516160-znlU (2026-07-27)
- Last verified: 2026-10-17

### Catch a 2-litre engine-oil overcharge on a car service bill with ChatGPT
- URL: https://www.realagentcases.com/use-cases/catch-a-2-litre-engine-oil-overcharge-on-a-car-service-bill-with-chatgpt/
- Outcome: Billed for a 6-litre pack of engine oil after a service on a Mahindra XUV 3XO at an authorised dealership, Gurpreet Singh Sodhi photographed the invoice and uploaded it to ChatGPT. It helped explain the invoice, identify the exact engine and its documentation, and prepare questions for the dealership. The dealership investigated, admitted the error, apologised and is re-issuing the invoice for 4 litres; the corrected bill is almost ₹1,500 lower.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/gurpreetsodhi_chatgpt-ai-mahindra-activity-7487071466249920513-Fry7 (2026-07-26)
- Last verified: 2026-10-17

### Spot a fake IRS letter as a scam by checking it with ChatGPT
- URL: https://www.realagentcases.com/use-cases/spot-a-fake-irs-letter-as-a-scam-by-checking-it-with-chatgpt/
- Outcome: When a letter claiming to be from the IRS arrived, Sara Wilbur's mother suspected a scam, checked it with ChatGPT and was right. Wilbur, who had encouraged her to trust her instincts, watch for red flags and use ChatGPT, shared it as an example of fraud awareness working.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Retold by the poster about their mother; the letter and ChatGPT's reply aren't shown.
- Source: https://www.linkedin.com/posts/saraasmith_my-mom-texted-me-a-few-minutes-ago-i-received-activity-7485527564183076865-Xtgi (2026-07-22)
- Last verified: 2026-10-17

### Visualize a furniture layout for an empty corner from a photo with ChatGPT
- URL: https://www.realagentcases.com/use-cases/visualize-a-furniture-layout-for-an-empty-corner-from-a-photo-with-chatgpt/
- Outcome: After three days lost in Pinterest boards, Tom's Guide's Kaycee Hill uploaded an unedited photo of an empty living-room corner to ChatGPT and described the chair, side table, bookcase and lamp she wanted. She then resized, removed and recoloured pieces in the generated images to test risky choices, and says she finished with a visual blueprint of what sizes to measure for and colours to buy, saving hundreds in potential returns.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/chatgpt/i-thought-ai-interior-design-was-a-gimmick-until-chatgpt-fixed-my-living-rooms-awkward-corner (2026-07-14)
- Last verified: 2026-10-17

### Get a dead AC condenser running again from photos with Grok
- URL: https://www.realagentcases.com/use-cases/get-a-dead-ac-condenser-running-again-from-photos-with-grok/
- Outcome: With their upstairs bedroom hitting 90°F, a poster asked Grok how to cool the house and found the outdoor condenser wasn't running. He says Grok walked him step by step through fuses, capacitors, contactors, multimeter use and Nest and furnace control-board wiring from his photos; after hours of near-misses he booked an HVAC visit, but Grok then spotted miswired Nest thermostat and control-board wires. After he fixed the wiring the AC ran, and he cancelled the appointment, estimating he saved $1,000+.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/EthanJones_45/status/2075451108139372626 (2026-07-10)
- Last verified: 2026-10-17

### Understand a formal Japanese notice from a new landlord and reply properly with ChatGPT
- URL: https://www.realagentcases.com/use-cases/understand-a-formal-japanese-notice-from-a-new-landlord-and-reply-properly-with/
- Outcome: A foreign resident in Osaka found a formal, indirect notice from his building's new owners. He photographed it and ChatGPT translated it, explained the implied meaning and flagged what to ask next; he checked the points with a real estate friend and emailed a formal Japanese reply. At the meeting, the owners asked tenants to leave and offered to cover moving costs and new-apartment fees, and he recorded and translated the meeting with Gemini and ChatGPT.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The landlord's offer was the owners' own decision; AI helped with translation and preparation.
- Source: https://www.linkedin.com/posts/adamgoulston_heres-how-ai-got-me-a-free-new-apartment-activity-7479749731716333568-aVqC (2026-07-06)
- Last verified: 2026-10-17

### File an Indian income tax return with ChatGPT as a cross-checker (it got some capital gains wrong)
- URL: https://www.realagentcases.com/use-cases/file-an-indian-income-tax-return-with-chatgpt-as-a-cross-checker-it-got-some-cap/
- Outcome: A filer in India used ChatGPT, with documents in a Project, as his main assistant for his income tax return. He says it was good at flagging common pitfalls and asking for supporting documents, and most valuable for checking calculations across documents. But it gave confident wrong guidance on REIT dividends and sales, misclassified some short- and long-term capital gains even with broker reports, and sometimes got totals wrong; he had to verify every important number. He'd still use it next year.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Useful cross-checker, but gave wrong REIT guidance, misclassified gains and miscalculated some totals.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ghitesh_this-week-i-filed-my-income-tax-return-itr-activity-7479806237682061312-sqEC (2026-07-06)
- Last verified: 2026-10-17

### Handle life admin after moving abroad with Claude: a mystery fridge alert, a tax-office letter and an insurance-app doctor chat
- URL: https://www.realagentcases.com/use-cases/handle-life-admin-after-moving-abroad-with-claude-a-mystery-fridge-alert-a-tax-o/
- Outcome: Nine months after moving a family of four from Connecticut to near Madrid, Rebecca Cretella uses Claude to handle daily life in a new language. In a few weeks it helped troubleshoot a flashing freezer indicator on the fridge in under a minute, chat with a doctor through the family's insurance app when a son came home with a hematoma, translate bloodwork results, and understand an audit letter from Spain's tax authority about a package a neighbour had sent.
- Agents: Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The medical items are from the writer's own account; this entry covers the admin side only and isn't medical advice.
- Source: https://www.businessinsider.com/kids-watch-me-use-ai-every-day-2026-7 (2026-07-04)
- Last verified: 2026-10-17

### Make a wedding seating chart from a guest spreadsheet with ChatGPT
- URL: https://www.realagentcases.com/use-cases/make-a-wedding-seating-chart-from-a-guest-spreadsheet-with-chatgpt/
- Outcome: Android Police's Olivia Locksley uploaded her guest spreadsheet (names, partners, relation to the couple) to ChatGPT with the number of tables and seats, general groupings and couples who must sit together. She then added nuance: spacing talkative and shy guests, seating a younger relative with his college friends and keeping feuding relatives apart. She says it handled the hard part, but the final draft still had to be checked and tweaked by hand.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/i-went-from-wedding-stress-to-wedded-bliss-with-chatgpt-as-my-wedding-planner/ (2026-06-30)
- Last verified: 2026-10-17

### A tax deduction strategy from ChatGPT that would have cost a business owner: an accountant's story
- URL: https://www.realagentcases.com/use-cases/a-tax-deduction-strategy-from-chatgpt-that-would-have-cost-a-business-owner-an-a/
- Outcome: An accountant says a client came in convinced ChatGPT had found a deduction strategy saving her $40,000 in taxes. He says the deduction category was misapplied for her business entity and the strategy relied on a rule changed by recent guidance; they caught it and filed correctly.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Per the accountant, the deduction was misapplied and based on outdated guidance; caught before filing.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: An accountant's account of a client's case; no documents shown.
- Source: https://www.linkedin.com/posts/joel-strong-3aba2270_a-client-called-me-last-week-convinced-ai-activity-7476062893801709568-0kvD (2026-06-26)
- Last verified: 2026-10-17

### Check a $930 dealership service estimate against warranty documents with Grok
- URL: https://www.realagentcases.com/use-cases/check-a-930-dealership-service-estimate-against-warranty-documents-with-grok/
- Outcome: In an interview summary, a truck owner told he'd been quoted $930 at the dealership for spark plug replacement. Before approving it, he uploaded his warranty and maintenance PDFs, the estimate and a screenshot and asked AI, 'Is any of this already covered under my warranty?' It found the work was already included in his prepaid maintenance package, saving $930. The interviewee's own post names Grok as the tool.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Retold in an interview post; the interviewee's own post names Grok.
- Source: https://www.linkedin.com/posts/sree-pradhip_aiinthecommunity-grok-activity-7474290619910270976-zXBh (2026-06-21)
- Last verified: 2026-10-17

### Keep a shared puppy feeding, potty and nap log in a ChatGPT group chat
- URL: https://www.realagentcases.com/use-cases/keep-a-shared-puppy-feeding-potty-and-nap-log-in-a-chatgpt-group-chat/
- Outcome: A couple with an 8-week-old puppy use one shared ChatGPT group chat to log when he eats, goes potty and naps, plus crate notes and 'is this normal?' questions, then ask ChatGPT to organise the pattern. The poster says it acts as shared memory, not a replacement for a vet or trainer.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written early on; no results reported yet.
- Source: https://www.linkedin.com/posts/brianjyip_new-puppy-hack-i-didnt-expect-a-shared-activity-7473777739746152448-DosT (2026-06-19)
- Last verified: 2026-10-17

### Send an email straight from ChatGPT through a connected Gmail account (attachments didn't work)
- URL: https://www.realagentcases.com/use-cases/send-an-email-straight-from-chatgpt-through-a-connected-gmail-account-attachment/
- Outcome: A TechRadar editor connected Gmail to ChatGPT and asked it to write and send an email with a story idea, background and a source link. ChatGPT showed the exact email and asked him to Allow or Deny; after he approved, it arrived seconds later. But it failed when he tried to add an attachment, and the feature only works on the web for paid plans. Gemini told him it couldn't send email directly.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Sending worked, but the attachment failed.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.techradar.com/ai-platforms-assistants/chatgpt/i-sent-an-email-without-opening-gmail-thanks-to-chatgpts-new-feature-then-i-found-the-catch (2026-06-09)
- Last verified: 2026-10-17

### Work out how to remove and repair a 1980s dishwasher from its model number with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-out-how-to-remove-and-repair-a-1980s-dishwasher-from-its-model-number-with/
- Outcome: A hired contractor couldn't work out how to remove the poster's grandparents' 1980s dishwasher, which was attached at the top. Hacker News user MattGaiser gave ChatGPT the model number; it explained why it wouldn't come out (a part had to be disconnected), found a replacement part and gave step-by-step instructions for taking that part out. They didn't end up buying the new part.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT explained the removal and found a part, but the family didn't buy the part, so the repair wasn't completed.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48423138 (2026-06-06)
- Last verified: 2026-10-17

### Build a weekly budget tracker in Google Sheets with Gemini
- URL: https://www.realagentcases.com/use-cases/build-a-weekly-budget-tracker-in-google-sheets-with-gemini/
- Outcome: A TechRadar guide shows how to prompt Gemini in Google Sheets to create a budget table with Category, Budgeted Amount, Actual Amount and Difference columns, then add a subtraction formula and conditional formatting so it updates as expenses are logged. It notes Gemini in Sheets needs a paid Google plan.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.techradar.com/ai-platforms-assistants/i-used-gemini-in-google-sheets-to-create-a-weekly-budget-that-updates-itself-heres-how (2026-06-04)
- Last verified: 2026-10-17

### Skip a likely upsell on a Jeep's $1,800 service estimate after checking it with ChatGPT
- URL: https://www.realagentcases.com/use-cases/skip-a-likely-upsell-on-a-jeep-s-1-800-service-estimate-after-checking-it-with-c/
- Outcome: At a Jeep dealership for a broken soft top, TikTok creator Debra was quoted $1,800 for an 80,000-mile service when she expected about $1,000. She asked for a printout, went outside and talked it through with ChatGPT, which told her the rear differential service was probably an upsell unless she towed heavy loads. She says skipping it saved $400. Motor1 notes the service is generally recommended every 30,000–60,000 miles and that experts say delaying it could cause problems later.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Second-hand: Motor1's write-up of the creator's TikTok video. Motor1 asked her for comment, but the article includes no reply.
- Source: https://www.motor1.com/news/796758/chatgpt-saves-money-car-dealership/ (2026-05-24)
- Last verified: 2026-10-17

### Recreate copycat Chick-fil-A nuggets with ChatGPT (it missed the breading)
- URL: https://www.realagentcases.com/use-cases/recreate-copycat-chick-fil-a-nuggets-with-chatgpt-it-missed-the-breading/
- Outcome: Tom's Guide's Amanda Caswell asked ChatGPT for the closest possible homemade Chick-fil-A nuggets recipe and used ChatGPT Voice to guide her while cooking; it kept saying she was on track. After one bite she realized it had skipped the most important part, the breading: top-rated copycat recipes specify things like Ritz crackers, sugar and baking soda. With no time to restart at about 8:15 p.m., she turned the chicken into a makeshift chicken parmesan.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The recipe gave no specifics on the breading, the key element, and cited no sources; the nuggets didn't taste right.
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-to-make-a-copycat-chick-fil-a-recipe-but-it-left-out-a-key-ingredient (2026-05-21)
- Last verified: 2026-10-17

### Add events from emails, voice notes and ticket screenshots to Google Calendar with Gemini
- URL: https://www.realagentcases.com/use-cases/add-events-from-emails-voice-notes-and-ticket-screenshots-to-google-calendar-wit/
- Outcome: An Android Police writer says Gemini fixed his main Google Calendar complaint, adding entries by hand. He asks Gemini by voice to schedule calls while out, uses it in Gmail to create events from invites, and sends it screenshots of concert-ticket presale times to make reminders. He notes it doesn't add guests automatically and the Gmail option doesn't always trigger on its own.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/i-almost-gave-up-on-google-calendar-until-this-gemini-hack-fixed-my-main-gripe/ (2026-05-17)
- Last verified: 2026-10-17

### Plan a date that breaks the dinner-and-a-movie routine with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-date-that-breaks-the-dinner-and-a-movie-routine-with-gemini/
- Outcome: Stuck for years in a 'dinner and a movie' routine, Android Police's Rahul Naskar asked Gemini to plan a different date. He says his partner felt the difference; telling stories it suggested brought a big smile to her face, and he has saved its idea of a 'stress-free' nature date for next time.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/used-gemini-to-plan-a-date-partner-noticed-the-difference/ (2026-05-14)
- Last verified: 2026-10-17

### Turn a messy weekly brain dump into a Google Calendar import file with Gemini
- URL: https://www.realagentcases.com/use-cases/turn-a-messy-weekly-brain-dump-into-a-google-calendar-import-file-with-gemini/
- Outcome: An Android Police writer plans the week each Sunday by writing out meetings, errands, bills and appointments in plain language and asking Gemini for a CSV to import into Google Calendar. The writer double-checks everything before importing, and says Gemini sometimes outputs plain text instead of a downloadable file, which they save as .csv themselves.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/paired-gemini-with-google-calendar-to-plan-my-week/ (2026-05-14)
- Last verified: 2026-10-17

### Replace paid reminder and to-do apps with Gemini scheduled actions
- URL: https://www.realagentcases.com/use-cases/replace-paid-reminder-and-to-do-apps-with-gemini-scheduled-actions/
- Outcome: An Android Police writer replaced IFTTT, Medisafe and Todoist with Gemini scheduled actions he already had through Google AI Pro: a weekday morning briefing, a Monday news digest, two daily medication reminders and a morning top-three task list. He cancelled about $13 a month in subscriptions, but notes Gemini only sends reminders (no dose tracking or interaction warnings) and that he relies on one company for everything, which makes him uneasy.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/geminis-scheduled-actions-replaced-apps-i-was-paying-for/ (2026-05-06)
- Last verified: 2026-10-17

### Rewrite weak resume bullets inside Google Docs with Gemini (then fix what it made up)
- URL: https://www.realagentcases.com/use-cases/rewrite-weak-resume-bullets-inside-google-docs-with-gemini-then-fix-what-it-made/
- Outcome: Android Police's Oluwaniyi Raji highlighted copied-job-description bullets in his resume and asked Gemini in Google Docs to rewrite them with action verbs and metric-led outcomes. The first output looked like it belonged to someone two levels above him, partly because Gemini made things up; generating took about 10 minutes and his own editing pass about 30 minutes before he had a resume he could submit to a recruiter.
- Agents: Gemini in Google Docs
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The first draft inflated his experience and included made-up details; it needed a 30-minute human editing pass.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/gemini-rewrote-my-resume-look-like-a-professional-on-paper/ (2026-05-05)
- Last verified: 2026-10-17

### Recover a RM 2,400 rental deposit in Kuala Lumpur using Malaysian tenancy rules found with Claude
- URL: https://www.realagentcases.com/use-cases/recover-a-rm-2-400-rental-deposit-in-kuala-lumpur-using-malaysian-tenancy-rules/
- Outcome: Two months after moving out, with no deposit back, a tenant in Kuala Lumpur gave Claude his lease, messages, move-out photos and repair receipts. He says it pointed to a 30-day return obligation, drafted a factual demand email citing the lease clauses, helped him challenge deductions (RM 1,750 claimed for floor damage versus an RM 250-400 market rate) and drafted a housing-board complaint. He recovered the full RM 2,400 in about 2 hours of work over three weeks.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/meetdebashis_ai-claudeai-productmanagement-activity-7456900421534797824-KZO_ (2026-05-04)
- Last verified: 2026-10-17

### Pick a new area, build a moving checklist and budget, and plan the interior with ChatGPT and Gemini
- URL: https://www.realagentcases.com/use-cases/pick-a-new-area-build-a-moving-checklist-and-budget-and-plan-the-interior-with-c/
- Outcome: A product manager says a friend moving house used ChatGPT to rank areas by commute, local shops and safety before choosing a place, then Gemini to build a moving checklist and rough budget in Google Sheets, then ChatGPT image generation for interior design ideas. The friend, who doesn't work in tech, said it was easier than expected.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Retold second-hand about a friend's move; no specifics or numbers.
- Source: https://www.linkedin.com/posts/gregoryambrose_chatgpt-and-i-helped-a-friend-move-house-activity-7455149537930842112-ueq9 (2026-04-29)
- Last verified: 2026-10-17

### Lower an electricity bill with ChatGPT's energy-saving tips
- URL: https://www.realagentcases.com/use-cases/lower-an-electricity-bill-with-chatgpt-s-energy-saving-tips/
- Outcome: Tom's Guide's Elton Jones told ChatGPT he paid a $230+ monthly energy bill and asked for tips to lower it. It pointed to repeated heater and air-conditioning use as a big driver; after changing habits such as fully turning off his consoles and installing LED bulbs, he says his Con Edison bill came to $130+ instead of $230+.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-to-cut-my-home-energy-use-and-these-tips-actually-saved-me-money (2026-04-23)
- Last verified: 2026-10-17

### Log spending by chatting and keep a budget workbook updated with a Claude Project
- URL: https://www.realagentcases.com/use-cases/log-spending-by-chatting-and-keep-a-budget-workbook-updated-with-a-claude-projec/
- Outcome: Budgeting apps never stuck for Mahnoor Faisal, so they had Claude build a full Excel workbook (transactions, live dashboard, monthly budget, savings goals, recurring bills, net worth) and set up a Claude Project with the template and instructions. Now they type things like the date and 'bought matcha, spent on commute', and Claude categorizes each transaction, asks when unsure (café or store?), updates the spreadsheet and returns the file, from phone or laptop. Faisal says it has done more for their spending than any budgeting app.
- Agents: Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The writer recreated the setup from scratch for the article; their real tracker isn't shown.
- Source: https://www.xda-developers.com/ditched-my-budgeting-app-for-claude-and-a-spreadsheet/ (2026-04-23)
- Last verified: 2026-10-17

### Repair a collision-damaged Honda Civic for under $1,000 instead of $10,000+ with Gemini as a mentor
- URL: https://www.realagentcases.com/use-cases/repair-a-collision-damaged-honda-civic-for-under-1-000-instead-of-10-000-with-ge/
- Outcome: After a collision left his daughter's 2019 Honda Civic with a $10,000+ shop quote and no collision coverage, a father spent weeks of evenings working with Gemini using photos, sensor data and questions. He says it helped him find a single crushed solenoid instead of a $4,000 transmission replacement, and guided which parts to buy and how to rebuild the front end. He reports a DIY total under $1,000 and the car running smoothly at 70+ mph.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/diegolruiz_google-geminiai-innovation-activity-7452691427672813568-A4np (2026-04-22)
- Last verified: 2026-10-17

### Search for a wedding venue with ChatGPT's Bridebook app (it couldn't beat his own search)
- URL: https://www.realagentcases.com/use-cases/search-for-a-wedding-venue-with-chatgpt-s-bridebook-app-it-couldn-t-beat-his-own/
- Outcome: A few months before his wedding, TechRadar's Hamish Hector tested Bridebook's ChatGPT integration with the venue brief he and his fiancée had used. Several venue costs were wrong compared with quotes they'd received, and he found it no more useful than scrolling Bridebook or Google, in some ways less, because its results were more restrictive.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Several listed costs were wrong versus real quotes, and the results were more restrictive than a normal search.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.techradar.com/ai-platforms-assistants/chatgpt/i-tried-a-new-chatgpt-tool-to-help-me-find-a-better-wedding-venue-and-im-very-glad-it-isnt-planning-my-big-day (2026-04-18)
- Last verified: 2026-10-17

### Stop a landlord's unannounced visits with a firm tenant-rights email drafted in Gemini
- URL: https://www.realagentcases.com/use-cases/stop-a-landlord-s-unannounced-visits-with-a-firm-tenant-rights-email-drafted-in/
- Outcome: After three unannounced visits from the landlord in one month, Android Police's Rahul Naskar asked Gemini for a calm but serious email that used tenant-rights language and mentioned escalation without sounding threatening. Gemini's draft cited Indian law on quiet enjoyment and asked for 24 hours' notice. The landlord replied that it was an oversight, promised to give 24 hours' notice, and there were no further unannounced visits for the rest of the stay.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/gemini-write-scary-email-to-landlord-and-it-got-the-job-done/ (2026-04-08)
- Last verified: 2026-10-17

### Clear a dishwasher error by cleaning the filter and resetting it with Grok
- URL: https://www.realagentcases.com/use-cases/clear-a-dishwasher-error-by-cleaning-the-filter-and-resetting-it-with-grok/
- Outcome: A poster says Google's AI told him a dishwasher error likely meant the machine needed repair or replacement, but Grok told him to check the filter, which was clogged, and gave exact instructions to reset it. It worked perfectly.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/haddypdaddy/status/2040925644645876104 (2026-04-05)
- Last verified: 2026-10-17

### Double-check a CPA-prepared draft tax return with ChatGPT and Claude and find two errors worth $2,686
- URL: https://www.realagentcases.com/use-cases/double-check-a-cpa-prepared-draft-tax-return-with-chatgpt-and-claude-and-find-tw/
- Outcome: Before filing, Katie Kormanik ran a CPA-prepared draft tax return and all of their 2025 tax documents through both ChatGPT and Claude. Both found the same two errors, which once corrected will save them $2,686.
- Agents: ChatGPT, Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Using two AI tools as a cross-check is the poster's approach. The two errors aren't described, and the $2,686 is the saving the poster expects once they're corrected, not a confirmed refund.
- Source: https://www.linkedin.com/posts/katiekormanik_ai-just-saved-me-and-my-husband-2686-in-activity-7446399497137672192-Pqwh (2026-04-05)
- Last verified: 2026-10-17

### Find target employers with Gemini Deep Research and prepare a son's job applications
- URL: https://www.realagentcases.com/use-cases/find-target-employers-with-gemini-deep-research-and-prepare-a-son-s-job-applicat/
- Outcome: A parent describes how the family got their son a job offer in under two months: Gemini Deep Research (several prompts) to build a target company list, a story-style resume, a master application, a unique cover letter per role using AI research, AI help finding hiring managers, and four AI-assisted interview prep sessions per interview. The poster reports about 20 applications, 5 interviews, 1 offer and 3 interview offers pending.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The offer wasn't final when this was posted.
- Source: https://www.linkedin.com/posts/justajobapp_activity-7445880028514496512-fuQr (2026-04-03)
- Last verified: 2026-10-17

### Sort a year of bank transactions for a freelancer's tax filing with Gemini, Claude and Lumo
- URL: https://www.realagentcases.com/use-cases/sort-a-year-of-bank-transactions-for-a-freelancer-s-tax-filing-with-gemini-claud/
- Outcome: XDA's Megan Ellis, a freelancer who files taxes three times a year, gave the same bank and credit-card exports and prompts to the free plans of Lumo, Gemini and Claude. Lumo missed a month of medication payments and about a third of the income. Gemini made no obvious errors and came within $100 of the manual income total. Claude hit its usage limit within two prompts at peak times but was the only one to match the manually filtered income total. Ellis still plans to do most of the taxes by hand.
- Agents: Gemini, Claude, Lumo
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Only Claude matched the manual income total; Lumo missed about a third of the income, and Gemini was close but not exact.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.xda-developers.com/using-gemini-claude-lumo-taxes/ (2026-03-30)
- Last verified: 2026-10-17

### Plan a twins' birthday party for 40 kids with ChatGPT and Claude head-to-head, from theme to RSVP tracking
- URL: https://www.realagentcases.com/use-cases/plan-a-twins-birthday-party-for-40-kids-with-chatgpt-and-claude-head-to-head-fro/
- Outcome: For their soon-to-be seven-year-old twins' party, Clifford Parker gave ChatGPT and Claude the same brief and budget over four rounds. Claude asked three questions first and produced a Minecraft sports day with a run sheet, prep list and budget tracker; ChatGPT found a local venue and entertainer with a phone number and booking script; ChatGPT's image generator won on invitations; and Claude, connected to Parker's computer, updated the RSVP spreadsheet from a phone message. It finished 2–2, with ChatGPT winning the tiebreaker.
- Agents: ChatGPT, Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test. Run by the parent on their own upcoming party.
- Source: https://www.linkedin.com/posts/cliffordparker_aitools-productivityforparents-chatgpt-activity-7444413237296578561-fzTo (2026-03-30)
- Last verified: 2026-10-17

### Check whether a senior dog's cardiac tests are covered by pet insurance using Claude Code
- URL: https://www.realagentcases.com/use-cases/check-whether-a-senior-dog-s-cardiac-tests-are-covered-by-pet-insurance-using-cl/
- Outcome: After a vet visit found a new heart murmur in his 13-year-old dog, a poster asked his always-on Claude Code assistant (via iMessage) to read the vet's records and estimate. He says it summarised the findings, explained each estimate line, found the insurance policy in years of email, worked out the tests would be 90% covered after the deductible, spotted that the vet might support direct pay, and sent the vet an email asking, copying his wife.
- Agents: Claude Code
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Uses the poster's own always-on Claude Code setup at home, which takes technical work to build.
- Source: https://www.linkedin.com/posts/tonymammoliti_claudecode-aiassistant-petinsurance-activity-7444029821891821568-Bgbn (2026-03-29)
- Last verified: 2026-10-17

### Spot forgotten trials and duplicate streaming plans with Gemini
- URL: https://www.realagentcases.com/use-cases/spot-forgotten-trials-and-duplicate-streaming-plans-with-gemini/
- Outcome: In a reply, a poster says Gemini saved him about $120/mo by pointing out a free trial he forgot about and that he had both Spotify and YouTube Premium, recommending a different credit card to earn points on his 2 biggest expenses, and suggesting alternatives to wasteful expenses.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/RustyOnLife/status/2037557116098027905 (2026-03-27)
- Last verified: 2026-10-17

### Track down a 1996 Volvo's stalling to a $35 fuel pump relay with Gemini
- URL: https://www.realagentcases.com/use-cases/track-down-a-1996-volvo-s-stalling-to-a-35-fuel-pump-relay-with-gemini/
- Outcome: A driver whose 1996 Volvo kept stalling, and whose two mechanics couldn't pin down why, spent about 15 minutes going back and forth with Gemini. It suggested a failed or failing fuel pump relay that acts up when the car isn't warmed up. He says the part cost $35 on eBay and plugs in with only a screwdriver needed to open a compartment.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/rminardi_recently-i-used-ai-to-diagnose-a-mysterious-activity-7443008542304964608-tm9b (2026-03-26)
- Last verified: 2026-10-17

### Lower a phone bill by $80 a month with a script from ChatGPT
- URL: https://www.realagentcases.com/use-cases/lower-a-phone-bill-by-80-a-month-with-a-script-from-chatgpt/
- Outcome: A poster uploaded his cell phone bill to ChatGPT and asked how it could save him money or whether another provider offered the same service for less. It gave him a script covering what he was overpaying for; he called T-Mobile, followed it and says his bill dropped by $80, about $960 a year, for under 5 minutes of his time.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/nathanginsbury_i-uploaded-my-cell-phone-bill-to-chat-gpt-activity-7442296515076956160-2xMd (2026-03-24)
- Last verified: 2026-10-17

### Find out where your dog is from doorbell camera clips with Gemini in Google Home
- URL: https://www.realagentcases.com/use-cases/find-out-where-your-dog-is-from-doorbell-camera-clips-with-gemini-in-google-home/
- Outcome: After months of vague 'Animal Detected' alerts from his Nest Doorbell, MakeUseOf's Bryan M. Wolfe asked Google Home (with Gemini) 'Hey Google, where is the dog?'. It answered in plain English with a clip: she was near the mailbox and had been there for about five minutes, which told him she wasn't wearing her boundary collar. The feature is part of Google Home Premium, from $10 a month.
- Agents: Gemini in Google Home
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.makeuseof.com/i-asked-gemini-where-my-dog-was-and-it-instantly-found-her/ (2026-03-20)
- Last verified: 2026-10-17

### Prepare a tiered question list for a specialist call with Claude
- URL: https://www.realagentcases.com/use-cases/prepare-a-tiered-question-list-for-a-specialist-call-with-claude/
- Outcome: Before a call with his wife's neuroendocrinologist, a poster gave Claude their clinical context and asked for a structured appointment prep. He says it organised questions into three tiers (what changes next steps, understanding the mechanism, and forward-looking), each with a note on why he's asking, so he could triage if time ran short. He shared a reusable template and says the goal is to be prepared, not to replace the doctor.
- Agents: Claude
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/arodriguez231_before-every-doctors-appointment-i-use-activity-7440045545819283456-O434 (2026-03-18)
- Last verified: 2026-10-17

### Lay out what was paid and owed in a vendor billing dispute with ChatGPT (it invented invoice numbers)
- URL: https://www.realagentcases.com/use-cases/lay-out-what-was-paid-and-owed-in-a-vendor-billing-dispute-with-chatgpt-it-inven/
- Outcome: A business owner fed ChatGPT all of the payment history, invoices and email threads from a vendor billing misunderstanding and asked it to set out exactly what had been paid and what was still owed. Five seconds before sending the email, he spot-checked a couple of invoice numbers and found they were all made up: they looked real and had the right format, but none existed. He didn't send it.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. ChatGPT invented every invoice number; caught by a spot-check just before sending.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/jamaalglenn_i-asked-chatgpt-to-help-me-settle-a-billing-activity-7437575639760666624-9NoV (2026-03-11)
- Last verified: 2026-10-17

### Find snack days and spirit days for each of three kids on school calendars with Gemini
- URL: https://www.realagentcases.com/use-cases/find-snack-days-and-spirit-days-for-each-of-three-kids-on-school-calendars-with/
- Outcome: A parent of three kids with three school calendars and many 'special events' photographs each calendar and gives it to Gemini with context, e.g. find his son's snack day (16 nut-free snacks) and all spirit days, filter to only that child's events, then add them to Google Calendar with a reminder to bring the snacks.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/mike-manzella-064a7537_three-kids-three-school-calendars-47-special-activity-7437318048509755392-hsAM (2026-03-11)
- Last verified: 2026-10-17

### Build a tiny house with no experience using ChatGPT for quantities, purchase order and cost-splitting, while catching frequent errors
- URL: https://www.realagentcases.com/use-cases/build-a-tiny-house-with-no-experience-using-chatgpt-for-quantities-purchase-orde/
- Outcome: With no building experience, Simone Solazzo and Anne Leijdekkers built a tiny home in the Netherlands over five months in 2025, asking ChatGPT questions constantly. It helped them think through deviations from the plan, do sums and measurements, work out quantities and buying order for around 250 items, and split costs between them. But they often doubted its guidance, and an architect friend showed its answers were frequently wrong; sometimes it stated false things when it didn't know. They relied on the friend and family for ventilation, insulation, electricity and water.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT helped with quantities, ordering and maths, but an architect friend found its answers were frequently wrong; the couple wouldn't rely on it alone.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/built-tiny-home-house-using-chatgpt-ai-useful-guidance-was-2026-3 (2026-03-07)
- Last verified: 2026-10-17

### Help a sixth grader learn to factor with Grok and ChatGPT as 'the greatest math teacher'
- URL: https://www.realagentcases.com/use-cases/help-a-sixth-grader-learn-to-factor-with-grok-and-chatgpt-as-the-greatest-math-t/
- Outcome: A parent photographed his daughter's sixth-grade worksheet and asked both Grok and ChatGPT to act as the greatest sixth-grade math teacher and explain to a 12-year-old how to factor each equation using the greatest common factor, with clear steps and ways to check her work. He says they both now know how to do it, and she did the rest of the worksheet without AI.
- Agents: Grok, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/matthewanelson_so-i-just-used-grok-and-chatgpt-to-help-me-activity-7435501231407517696-aCy7 (2026-03-06)
- Last verified: 2026-10-17

### Find a failed AC contactor and fix it for $15 with Claude's step-by-step troubleshooting
- URL: https://www.realagentcases.com/use-cases/find-a-failed-ac-contactor-and-fix-it-for-15-with-claude-s-step-by-step-troubles/
- Outcome: When his thermostat rebooted every time the AC tried to start, a homeowner worked through it with Claude: checking thermostat wiring, then the air handler board and fuse, then isolating the cooling circuit, which pointed to the outdoor unit. There he found a contactor with degraded coil wiring, ordered a $15 replacement and swapped it in 20 minutes. He says the AC runs perfectly, versus a $150 minimum service call he expected to reach $300-400.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/brandonpenrod_my-ac-died-last-week-thermostat-kept-rebooting-activity-7435313622719012865-XmAV (2026-03-05)
- Last verified: 2026-10-17

### Re-learn high school maths with quizzes and a study guide in Gemini Guided Learning
- URL: https://www.realagentcases.com/use-cases/re-learn-high-school-maths-with-quizzes-and-a-study-guide-in-gemini-guided-learn/
- Outcome: An XDA writer used Gemini's free Guided Learning mode to retake high school maths and history after two decades: it built 10-question multiple-choice quizzes, explained wrong answers, reviewed his weak spots and produced a study guide he turned into a Google Doc. He says a few sessions rebuilt his knowledge, and he also used it to learn food-ordering phrases for Paris, but he doesn't see it replacing a teacher.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.xda-developers.com/i-used-google-gemini-to-retake-high-school-after-two-decades-and-heres-how-it-went/ (2026-02-24)
- Last verified: 2026-10-17

### Find extra tax refund money with Grok's help
- URL: https://www.realagentcases.com/use-cases/find-extra-tax-refund-money-with-grok-s-help/
- Outcome: A poster says he asked Grok to help with his taxes and it helped him get around $1,500 more on his tax return, adding that it also helped him learn about taxes he was never taught.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Patrick15212115/status/2025658707502391506 (2026-02-22)
- Last verified: 2026-10-17

### Get a dog with sudden foamy coughing to an emergency vet in time after ChatGPT flagged possible bloat
- URL: https://www.realagentcases.com/use-cases/get-a-dog-with-sudden-foamy-coughing-to-an-emergency-vet-in-time-after-chatgpt-f/
- Outcome: When the family dog began coughing up foam after drinking, Chase Williams asked ChatGPT to question them like a vet would. After asking about breed, age, frequency and when it started, ChatGPT said it could be urgent and might be bloat, which in bigger dogs can be fatal within hours. The dog was put in the car immediately; the vet diagnosed bloat needing emergency surgery and said the dog had arrived just in time. Williams stresses ChatGPT is not a vet and shouldn't be trusted on its own.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/chaseawilliams_i-saved-my-dogs-life-using-chatgpt-which-activity-7429517520015085568-cvAp (2026-02-17)
- Last verified: 2026-10-17

### Make sense of dealership lease printouts mid-visit with ChatGPT
- URL: https://www.realagentcases.com/use-cases/make-sense-of-dealership-lease-printouts-mid-visit-with-chatgpt/
- Outcome: A first-time car shopper who found the dealer's math confusing asked for a minute alone, took screenshots of the printouts, gave ChatGPT some context and asked how to negotiate and which options were right for her family. She says it explained the pros and cons clearly and told her what to negotiate; she followed about 90% of its advice, said 'no' about 4 times to 'meet in the middle' offers, and leased the car on her terms.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/elainezelby_chatgpt-helped-me-buy-a-car-this-weekend-activity-7429215851125415936-FGcW (2026-02-16)
- Last verified: 2026-10-17

### Plan a fun, free weekend with a 3-year-old with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-fun-free-weekend-with-a-3-year-old-with-gemini/
- Outcome: Away for Canada's February long weekend, Android Police's Jade Lassalle wanted a fun-filled plan for her partner and 3-year-old daughter. 'Help me plan my weekend' got clarifying questions and vibe options; 'Help me plan a fun weekend with my 3-year-old' gave much better results. She found Google Maps searches for things to do in Hamilton showed mostly restaurants and shopping, not the fun and free ideas she wanted.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/i-let-gemini-plan-my-weekend/ (2026-02-15)
- Last verified: 2026-10-17

### Cut a $195,628 uninsured ER bill to about $32,500 by checking it against Medicare rules with Claude
- URL: https://www.realagentcases.com/use-cases/cut-a-195-628-uninsured-er-bill-to-about-32-500-by-checking-it-against-medicare/
- Outcome: After a relative died following four hours of emergency care, Matt Rosenberg got the hospital's coded UB-04 form and asked Claude to compare each charge with what Medicare would pay. Claude spotted unbundled stent supplies worth over $100,000, a bypass charge for surgery that never happened, and ventilation billed alongside critical care, and estimated Medicare would have paid about $28,675. Rosenberg had ChatGPT check the analysis, spot-checked the rules, and sent a six-page letter. The hospital countered at $36,356 and they settled at roughly $32,500.
- Agents: Claude, ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Rosenberg says they cross-checked Claude's figures with ChatGPT and the Medicare rules themselves, and doesn't recommend using AI as a lawyer.
- Source: https://www.businessinsider.com/claude-helped-negotiate-hospital-bill-discount-medicare-ai-assistant-2026-2 (2026-02-13)
- Last verified: 2026-10-17

### Load a child's soccer schedule onto the family calendar with 2-hour blocks using Gemini
- URL: https://www.realagentcases.com/use-cases/load-a-child-s-soccer-schedule-onto-the-family-calendar-with-2-hour-blocks-using/
- Outcome: A parent screenshotted his son's soccer schedule, uploaded it to Gemini, asked it to create events with a 2-hour window and told it which family calendar to use. He says two minutes later every game was on the family calendar, saving about 15 minutes of manual entry.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/lancetmoore_i-uploaded-my-sons-soccer-schedule-to-gemini-activity-7427773527983808512-1kxI (2026-02-12)
- Last verified: 2026-10-17

### Fix a washing machine that won't fill using Gemini, after ChatGPT picked the wrong part
- URL: https://www.realagentcases.com/use-cases/fix-a-washing-machine-that-won-t-fill-using-gemini-after-chatgpt-picked-the-wron/
- Outcome: A poster described his broken washing machine to Gemini, which helped him run diagnostics, found it wasn't taking in water and pointed to the inlet valve. He used ChatGPT's shopping help to find and order the part, but the first part didn't fit even though ChatGPT said it would. He installed the right one with Gemini's guidance in about 1.5 hours total, saving about $120 on labour.
- Agents: Gemini, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mateusz-nowotka-34aa1018a_gemini-chatgpt-ai-activity-7427437526329479168-KG3W (2026-02-11)
- Last verified: 2026-10-17

### Plan a week of less starch-heavy meals around local foods and budget with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-week-of-less-starch-heavy-meals-around-local-foods-and-budget-with-gemini/
- Outcome: Android Police's Irene Okpanachi, whose diet was mostly starch, asked Gemini to act as a nutrition coach, held its questions until a 'READY' safe word, and described habits, local foods and budget. After revisions to swap out foods that weren't affordable or easy to find, Gemini's plan moved most starch to lunch, kept breakfast and dinner lighter and more protein-forward, and set simple rules such as a fist-sized starch portion and eating vegetables and protein first.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/meal-planning-with-gemini/ (2026-02-11)
- Last verified: 2026-10-17

### Build a biweekly grocery list on a $175 budget with Gemini
- URL: https://www.realagentcases.com/use-cases/build-a-biweekly-grocery-list-on-a-175-budget-with-gemini/
- Outcome: Tom's Guide's Elton Jones gave Gemini and ChatGPT the same prompt: a $175 biweekly grocery budget, the foods he usually buys, and a request for a categorized, priced table. Gemini said $175 over two weeks was realistic if he balanced splurges like steak and shrimp with rice and potatoes, and gave a table by category, item, quantity and estimated cost with a total. ChatGPT's similar chart added items he doesn't usually buy, like eggs and shredded cheese. He picked Gemini as the better financial adviser overall.
- Agents: Gemini, ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/i-tested-gemini-vs-chatgpt-for-money-management-heres-which-one-helped-me-the-most (2026-02-09)
- Last verified: 2026-10-17

### Use Gemini as a job-search co-worker after a layoff: resume, cover letters, outreach and interview prep
- URL: https://www.realagentcases.com/use-cases/use-gemini-as-a-job-search-co-worker-after-a-layoff-resume-cover-letters-outreac/
- Outcome: The day his role was eliminated, a poster started using Gemini as a full-time job-search co-worker. Over three weeks and more than 40 hours, he says it helped him optimise his LinkedIn profile, update his resume, write custom cover letters, networking messages and recruiter follow-ups, and prepare for interviews. He reviews and challenges everything it produces.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: No job outcome reported yet.
- Source: https://www.linkedin.com/posts/jeffpalen_the-day-my-role-was-eliminated-i-told-google-activity-7425206642947850240-NxRA (2026-02-05)
- Last verified: 2026-10-17

### Turn a school-year calendar into draft Google Calendar events with Gemini
- URL: https://www.realagentcases.com/use-cases/turn-a-school-year-calendar-into-draft-google-calendar-events-with-gemini/
- Outcome: Instead of typing in the kids' school calendar for the next school year, Joshua Kaan gave Gemini the details. Gemini drafted the events and Kaan approved them. The one thing it couldn't do was add other people's email addresses as guests on the events.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. Gemini drafted and added the events after approval, but couldn't add other email addresses as guests.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/joshuakaan_time-saver-this-weekend-instead-activity-7424142060770865152-f0h- (2026-02-02)
- Last verified: 2026-10-17

### Appeal a denied cardiac test using studies found with ChatGPT Deep Research
- URL: https://www.realagentcases.com/use-cases/appeal-a-denied-cardiac-test-using-studies-found-with-chatgpt-deep-research/
- Outcome: When insurance denied the poster's husband's cardiologist-ordered nuclear stress test as 'experimental' (a $1,500 out-of-pocket cost), the poster used ChatGPT Deep Research to find peer-reviewed studies on the test's effectiveness: five studies in ten minutes. Five minutes later they had a draft appeal letter for the cardiologist to review and send. The poster says a week later the insurer reversed its decision and covered the test.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/njshelsby_insurance-denied-my-husbands-heart-test-activity-7421932830110343169-JPiY (2026-01-27)
- Last verified: 2026-10-17

### Tailor a resume, cover letter and LinkedIn summary with ChatGPT, then get Gemini to judge them
- URL: https://www.realagentcases.com/use-cases/tailor-a-resume-cover-letter-and-linkedin-summary-with-chatgpt-then-get-gemini-t/
- Outcome: A writer fed ChatGPT a job description and his existing application details to produce a tailored resume (two attempts; the first looked too plain), a one-page cover letter and a LinkedIn summary with recruiter keywords. He then asked Gemini to play the hiring manager: it scored him 8.5 out of 10 with 'a very high probability of landing an interview', but said the documents were too 'tech-first' rather than 'finance-first' for this employer.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A practice application; no real hiring result.
- Source: https://www.tomsguide.com/ai/i-had-chatgpt-write-my-resume-linkedin-summary-and-cover-letter-then-asked-gemini-if-i-would-get-the-job (2026-01-23)
- Last verified: 2026-10-17

### Narrow down why a gas dryer takes three hours with a ChatGPT video chat before calling a repair company
- URL: https://www.realagentcases.com/use-cases/narrow-down-why-a-gas-dryer-takes-three-hours-with-a-chatgpt-video-chat-before-c/
- Outcome: When a dryer took three hours to dry a load (normally 45 minutes), Lisa-Anne Chung showed ChatGPT the model number, the parts that weren't working and the pipes at the back over a video chat. It identified the model from a photo, walked Chung through checks such as whether the burner glowed and listening for clicks, gave a shortlist of likely causes, and suggested Sunnyvale repair companies via Yelp with a fair diagnostic fee. Chung says they had a solid hypothesis in about 10 minutes and could talk knowledgeably when they called a repair company.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://lisaslab.substack.com/p/broken-dryer-chatgpt (2026-01-22)
- Last verified: 2026-10-17

### Diagnose a broken garage door opener and replace the carriage for $22.48 with Gemini
- URL: https://www.realagentcases.com/use-cases/diagnose-a-broken-garage-door-opener-and-replace-the-carriage-for-22-48-with-gem/
- Outcome: A homeowner described his garage door opener's behaviour to Gemini, which identified a failed carriage, pointed him to where to buy a new one and found a DIY installation video. He says he spent ten minutes with Gemini and $22.48, versus a service call and repair he estimates would have cost several hundred dollars.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ajklingenmaier_my-latest-fascination-with-ai-is-using-it-activity-7419872320879755264-EAr1 (2026-01-21)
- Last verified: 2026-10-17

### Build a shared household inventory of bills, policies and subscriptions from Gmail with a Gemini prompt
- URL: https://www.realagentcases.com/use-cases/build-a-shared-household-inventory-of-bills-policies-and-subscriptions-from-gmai/
- Outcome: Nick Malham shares a Gemini prompt for turning six months of Gmail into one shared Google Sheet of household admin: insurances, banking and home loan, utilities, car cover, phone and internet, subscriptions, and will and power of attorney, with company, cover type, who it's for, account number, cost and frequency, payment type and key phone numbers. Malham suggests connecting Gemini to Google Workspace first, and checking a 12-month bank export for anything that never reached the inbox.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task. Account numbers and policy details are sensitive; keep the shared sheet private to the household.
- Source: https://www.linkedin.com/posts/nickmalham_ive-been-playing-with-gemini-recently-trying-activity-7419222178426175489-u_l4 (2026-01-20)
- Last verified: 2026-10-17

### Figure out why a car suddenly won't go over 19 mph from a dashboard photo with ChatGPT
- URL: https://www.realagentcases.com/use-cases/figure-out-why-a-car-suddenly-won-t-go-over-19-mph-from-a-dashboard-photo-with-c/
- Outcome: On a dark, rainy drive, TechRadar's Graham Barlow's second-hand car would only crawl and the dash showed '19 mph limit'. He photographed the dashboard and asked ChatGPT; it said a speed limiter was switched on at 19 mph and was probably controlled by a button on a stalk behind the wheel. He found the button labelled 'Limit', turned it off and drove on without calling roadside recovery.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.techradar.com/ai-platforms-assistants/chatgpt/chatgpt-saved-me-from-a-roadside-nightmare-this-is-exactly-what-we-should-be-using-ai-for-not-for-replacing-creatives (2026-01-15)
- Last verified: 2026-10-17

### Pick a breed, name, shopping list and care plan for a first family puppy with Gemini
- URL: https://www.realagentcases.com/use-cases/pick-a-breed-name-shopping-list-and-care-plan-for-a-first-family-puppy-with-gemi/
- Outcome: Planning a first dog with their son, Allison Marx used mostly Gemini to recommend breeds that suited their lifestyle and requirements (they chose a Havanese), come up with a name (Chip), build a shopping list and draft a care plan. Two weeks after bringing him home, Marx says the family couldn't be happier.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/allisonsmarx_ai-generativeai-gemini-activity-7416937578337312768-4hl4 (2026-01-13)
- Last verified: 2026-10-17

### Check a 'take-home task' repo from a job offer for malware with ChatGPT before running it
- URL: https://www.realagentcases.com/use-cases/check-a-take-home-task-repo-from-a-job-offer-for-malware-with-chatgpt-before-run/
- Outcome: A developer offered a 'unicorn' job taking over a software project was asked to review and run the company's GitHub repository. Instead of cloning or running it, he asked ChatGPT to analyze it, and says it found hidden malware in what looked like a harmless library file, built to connect to an outside server and steal files, keys and credentials when the project was started. He calls it a common scam aimed at job seekers.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/makowskid/status/2010206469069709692 (2026-01-11)
- Last verified: 2026-10-17

### Clear a car's no-start service code caused by a slightly open window with Grok
- URL: https://www.realagentcases.com/use-cases/clear-a-car-s-no-start-service-code-caused-by-a-slightly-open-window-with-grok/
- Outcome: After leaving the driver's window about 1/4" open for roughly 36 hours, a poster's car wouldn't start and they jump-started it with a portable charger; the service manual pointed them to the dealer. The poster says Grok quickly found there's a dash service code for the driver window and the fix was to roll the window down and back up, which cleared the code and saved a dealer trip.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/walkshappening/status/2009330737778975205 (2026-01-08)
- Last verified: 2026-10-17

### Build an interactive kids' chore tracker that unlocks weekend screen time with Gemini
- URL: https://www.realagentcases.com/use-cases/build-an-interactive-kids-chore-tracker-that-unlocks-weekend-screen-time-with-ge/
- Outcome: After two weeks of the kids home for the holidays, Liam Darmody used Gemini to create a master tracking sheet for chores, replacing fridge lists that never stuck. Each day a child completes all tasks, a green progress bar unlocks 20% of their weekend iPad time; hitting all five days unlocks a badge and all of their weekend iPad time. Gemini worked out formulas Darmody says were more complex than anything they could have written, and also gave a prompt for prototyping an app.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/liamdarmody1_after-14-days-of-kids-being-home-for-the-activity-7413647106676072448-81H9 (2026-01-04)
- Last verified: 2026-10-17

### Fix a dead freezer with a £10 wire after Grok's voice-and-camera mode found a known fault
- URL: https://www.realagentcases.com/use-cases/fix-a-dead-freezer-with-a-10-wire-after-grok-s-voice-and-camera-mode-found-a-kno/
- Outcome: When a freezer stopped working, Marc Landers gave Grok the make, model and problem. Grok had them turn on voice and camera mode, open the door, remove the top hinge cover (no tools needed) and point the camera at a wire it said was a known fault for that model. Grok gave the price and an Amazon link for the part; the £10 wire arrived next day and took five minutes to replace, instead of paying a repair technician.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/marclanders/status/2003817594126520537 (2025-12-24)
- Last verified: 2026-10-17

### Get a faulty demo car replaced under South Africa's Consumer Protection Act with Grok's help
- URL: https://www.realagentcases.com/use-cases/get-a-faulty-demo-car-replaced-under-south-africa-s-consumer-protection-act-with/
- Outcome: Two days after buying a demo Chery Tiggo 4 Pro, the tyre-pressure monitoring system failed at 120 km/h. The poster (@viking_boer) used South Africa's Consumer Protection Act and Grok's step-by-step advice to reject the car. The dealership resisted, but within two weeks the poster got a brand-new replacement, with the leather seats swapped over.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/viking_boer/status/2002453061718139190 (2025-12-20)
- Last verified: 2026-10-17

### Compare pet insurance policies and find a promo code with Grok, saving about $1,000 a year
- URL: https://www.realagentcases.com/use-cases/compare-pet-insurance-policies-and-find-a-promo-code-with-grok-saving-about-1-00/
- Outcome: After their golden retriever's insurance rose to $182/month, a dog owner got a $110/month quote elsewhere and uploaded both policies, their disclosure statements and the quote to Grok for a detailed comparison. In 24 seconds it produced a table of key inclusions, which was more comprehensive, benefits and drawbacks of switching, and a recommendation to switch. The owner turned down three retention offers, cancelled, then had Grok search for public promotions and found one that saved a further 20%. They say they got more cover and will save about $1,000 a year, in about 15 minutes.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/jordan-williams-land_i-used-ai-to-help-me-save-1000year-on-activity-7404509068964425731-w6sb (2025-12-10)
- Last verified: 2026-10-17

### Trace a Mitsubishi engine light to a £100 oxygen sensor instead of a £2,500 catalytic converter with Grok
- URL: https://www.realagentcases.com/use-cases/trace-a-mitsubishi-engine-light-to-a-100-oxygen-sensor-instead-of-a-2-500-cataly/
- Outcome: After independent garages kept resetting the engine management light and a main dealer quoted £2,500 for a new catalytic converter, Paul Ryan bought a £20 OBD reader, downloaded the data and uploaded it to Grok. Grok said it was the post-cat oxygen sensor, not the converter. A mechanic agreed to replace just the sensor (£100 part, £100 labour) once Ryan signed that it might not fix it. The light stayed off and the sensors read normal; Ryan says it saved £2,000.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/pauljryan_diagnosing-an-engine-fault-with-ai-my-mitsubishi-activity-7402796531117948928-SNlw (2025-12-05)
- Last verified: 2026-10-17

### Get a wrongly coded medical bill corrected with Grok
- URL: https://www.realagentcases.com/use-cases/get-a-wrongly-coded-medical-bill-corrected-with-grok/
- Outcome: After weeks of getting the runaround from both her doctor's office and her insurer, a poster put all the bill information into Grok. She says it found the coding issue and told her how to press it with her doctor, and she avoided a huge medical bill.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/BamaGlamaGirl/status/1996705075574747296 (2025-12-04)
- Last verified: 2026-10-17

### Find out a leaking shower cartridge is under lifetime warranty with Grok
- URL: https://www.realagentcases.com/use-cases/find-out-a-leaking-shower-cartridge-is-under-lifetime-warranty-with-grok/
- Outcome: A Delta shower installed in 2009 was leaking badly from its cartridge. The poster (@JustAnAvgJoeW) asked Grok, which said the faucet had a lifetime warranty; they ordered the replacement part the same day and planned to fit it themselves after a video, which they estimated would save $400–$500 over a plumber.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. The part was ordered under warranty, but the post was written before it arrived or was fitted; the $400–$500 saving is the poster's estimate.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/JustAnAvgJoeW/status/1995703807380955627 (2025-12-02)
- Last verified: 2026-10-17

### Build a month's family meal plan around kids' preferences, three stores and a custody schedule with Grok Projects
- URL: https://www.realagentcases.com/use-cases/build-a-month-s-family-meal-plan-around-kids-preferences-three-stores-and-a-cust/
- Outcome: Talking rather than typing, the poster behind @reasonablTchDad gave Grok Projects three kids' food preferences, ten to fifteen past meal ideas and who liked which, December school-vacation days, which items are available at which of three stores, and the custody schedule. Grok produced a 'kids working doc' with preference and meal tables, and a December plan with daily meals, a shopping list split by store, holiday snacks and easy lunches for the two weeks the kids are home, plus a Notion-ready version. Both docs stay in the project for next month's plan.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/reasonablTchDad/status/1995653256769397107 (2025-12-02)
- Last verified: 2026-10-17

### Fix a diesel truck's limp-mode problem part by part over six weeks with Grok and YouTube
- URL: https://www.realagentcases.com/use-cases/fix-a-diesel-truck-s-limp-mode-problem-part-by-part-over-six-weeks-with-grok-and/
- Outcome: After a dealer charged $7,000 for a new catalytic converter and the GMC diesel still went into limp mode while towing an Airstream, Greg Wall spent six weeks talking to Grok and watching YouTube. Grok helped find part numbers, estimate costs and rate the difficulty of each job; Wall replaced the #2 exhaust gas temperature sensor, alternator, #9 fuel injector, intercooler pipe and air intake box, filter and pipe, but couldn't loosen the NOx sensor nut. The couple bought a new truck anyway; on a later 2-hour drive the diesel showed no warnings and set a 21.2 mpg record.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Wall couldn't replace the NOx sensor and the couple bought a new truck before the job was finished; 'fixed' is the poster's own judgement after one long drive.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://airstreamtime.com/2025/12/01/grok-youtube-and-i-fixed-the-gmc/ (2025-12-01)
- Last verified: 2026-10-17

### Get a car with a loose hose and a flooded engine started again with Gemini
- URL: https://www.realagentcases.com/use-cases/get-a-car-with-a-loose-hose-and-a-flooded-engine-started-again-with-gemini/
- Outcome: When his car wouldn't start, a poster sent Gemini a photo of a loose hose under the hood. It identified the air intake; he reattached it but the car still wouldn't start, and from the symptoms Gemini said the engine was likely flooded and explained how to clear it (gas pedal down, turn the key). The engine started, saving him a tow; he still booked a professional check.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/iermilov_ai-gemini-vibecoding-activity-7400867144412061696-uObt (2025-11-30)
- Last verified: 2026-10-17

### Set up a weekly meal plan that ChatGPT agent generates every Monday
- URL: https://www.realagentcases.com/use-cases/set-up-a-weekly-meal-plan-that-chatgpt-agent-generates-every-monday/
- Outcome: Tom's Guide's Christoph Schwaiger shares a prompt that schedules ChatGPT agent to generate a weekly meal plan every Monday at 08:00 from food bloggers, local restaurant menus and trending recipes, choosing recipes that share key ingredients to minimize shopping, with optional extras like a grocery list, Google Sheets export or nutrition overview. It's a setup guide and doesn't report a finished week.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.tomsguide.com/ai/chatgpt/i-let-chatgpt-agent-find-recipes-for-me-and-it-was-surprisingly-good (2025-11-29)
- Last verified: 2026-10-17

### Find a check-engine air leak a mechanic missed with Grok and ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-a-check-engine-air-leak-a-mechanic-missed-with-grok-and-chatgpt/
- Outcome: After their mechanic swapped 3 parts without fixing a check engine light and was about to replace another, a driver stopped the mechanic and used ChatGPT and Grok instead. The driver says they found the air leak on their own and fixed the light.
- Agents: Grok, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/always_improve7/status/1994916407587606672 (2025-11-29)
- Last verified: 2026-10-17

### Turn a handwritten family pie recipe into an infographic with Gemini and ChatGPT (it garbled the recipe)
- URL: https://www.realagentcases.com/use-cases/turn-a-handwritten-family-pie-recipe-into-an-infographic-with-gemini-and-chatgpt/
- Outcome: A TechRadar writer photographed his mother's 30-year-old handwritten apple pie recipe and asked Gemini and ChatGPT to turn it into a clear, cartoon-style infographic. Gemini's version was cute but full of misspellings and nonsense like 'Dash nutamon' and a cheese grater, and corrections didn't fix it; ChatGPT did marginally better but left out details. Claude gave the clearest, if plain, recreation. He called Gemini's first try unusable and kept his paper original.
- Agents: Gemini, ChatGPT, Claude
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Gemini and ChatGPT misread the handwriting and invented words and steps; repeated corrections didn't fix it.
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.techradar.com/ai-platforms-assistants/sorry-chatgpt-and-gemini-cant-improve-on-my-moms-handwritten-apple-pie-recipe (2025-11-29)
- Last verified: 2026-10-17

### Find and book a restaurant table that fits gluten-free and veggie diets with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-and-book-a-restaurant-table-that-fits-gluten-free-and-veggie-diets-with-cha/
- Outcome: Tom's Guide's Alex Hughes asked ChatGPT agent to find and book a table in Bristol for next Wednesday between 7 and 8 pm with vegetarian options at about £30-40 a head. It asked about diets and party size (3 people, one gluten-free) and about two minutes later returned three options, all with tables free at that time and with vegetarian and gluten-free choices, each with an explanation; once he picked one, he only had to add contact details and confirm the booking.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/chatgpt/i-use-chatgpt-to-do-all-of-my-restaurant-bookings-and-it-saves-me-so-much-time (2025-11-16)
- Last verified: 2026-10-17

### Winterize old plumbing from photos with Gemini, and know when to call a plumber
- URL: https://www.realagentcases.com/use-cases/winterize-old-plumbing-from-photos-with-gemini-and-know-when-to-call-a-plumber/
- Outcome: At his late mother's 100+ year old house, a poster uploaded photos of three plumbing issues to Gemini Pro. For two (valves for outside hose bibs and a shower) it identified the plumbing type and explained how to winterize, which he did without a hitch; for a leaky hot-water boiler it diagnosed the problem and told him to call a plumber, which he did.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/neddesmond_i-just-used-google-gemini-pro-to-solve-three-activity-7394125364752670721-WESf (2025-11-11)
- Last verified: 2026-10-17

### Turn a set of favourite recipes into a week's meal plan, weekend baking schedule and aisle-sorted grocery list with Grok
- URL: https://www.realagentcases.com/use-cases/turn-a-set-of-favourite-recipes-into-a-week-s-meal-plan-weekend-baking-schedule/
- Outcome: To feel less stressed for the week, the poster (@FeralHWife85) gave Grok five dinner recipes, a potato bun recipe, a French bread recipe and three sourdough recipes and asked it to plan the week's meals and the weekend around them. The poster says it did so flawlessly and produced a grocery list organized by store department.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/FeralHWife85/status/1986907373848633751 (2025-11-07)
- Last verified: 2026-10-17

### Bring down an extended car warranty quote from $22,600 to $7,400 with ChatGPT's advice
- URL: https://www.realagentcases.com/use-cases/bring-down-an-extended-car-warranty-quote-from-22-600-to-7-400-with-chatgpt-s-ad/
- Outcome: When his car warranty expired, a driver was quoted $22,600 by the dealership for a five-year, 60,000-mile extension. He says ChatGPT told him that was far too high; the dealer dropped the same warranty to $11,800 within five minutes. Still told it was too much, he called another dealership and was offered $7,400 for seven years and 90,000 miles.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/david-bodnick-25a7791_ai-artificialintelligence-negotiation-activity-7391142851226681344-cXNI (2025-11-03)
- Last verified: 2026-10-17

### Overturn a pet insurer's pre-existing-condition denial with a formal ChatGPT-drafted demand letter
- URL: https://www.realagentcases.com/use-cases/overturn-a-pet-insurer-s-pre-existing-condition-denial-with-a-formal-chatgpt-dra/
- Outcome: A pet insurer denied a claim for a dog's root canal, citing a minor cosmetic chip in the same tooth years earlier, before coverage. Hacker News user awillen used ChatGPT to write a formal demand letter citing California rules on what counts as a pre-existing condition, plus case law on insurers' burden when invoking exclusions; the poster checked the cases were real. The letter gave 14 days before filing in small claims court, and on day 14 the insurer said it would pay in full.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=45735660 (2025-10-28)
- Last verified: 2026-10-17

### Get a GoDaddy business email charge refunded with ChatGPT's guidance
- URL: https://www.realagentcases.com/use-cases/get-a-godaddy-business-email-charge-refunded-with-chatgpt-s-guidance/
- Outcome: A poster charged $83.88 for GoDaddy business email followed ChatGPT's guidance on asking for a refund. The poster says GoDaddy refunded the full amount and moved the account to a $35.88 discounted plan, which they count as a $48 saving.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/WorksBreaks/status/1982112202233377114 (2025-10-25)
- Last verified: 2026-10-17

### Fight off a £5,800 landlord claim in court using Grok and ChatGPT
- URL: https://www.realagentcases.com/use-cases/fight-off-a-5-800-landlord-claim-in-court-using-grok-and-chatgpt/
- Outcome: A UK poster says a former landlord tried to sue them for £5,800 in a year-long dispute. They used Grok and ChatGPT for the whole process from start to finish instead of paying a lawyer's quoted £3,000, and say they won and received the court's closure letter. They add that this is not legal advice.
- Agents: Grok, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Jesusnumba1fan/status/1976050282455847274 (2025-10-08)
- Last verified: 2026-10-17

### Create a Ninjago-themed birthday quest with activity scroll and medallions using Gemini and ChatGPT
- URL: https://www.realagentcases.com/use-cases/create-a-ninjago-themed-birthday-quest-with-activity-scroll-and-medallions-using/
- Outcome: A parent who once ran a kids' party business used Gemini and ChatGPT to create visual assets for their son's 7th birthday quest: a ninja 'scroll' showing all the activities and an elemental medallion that kids filled with stickers as they completed challenges. The parent says the scroll helped walk the kids through the activities and the medallions kept them focused and became souvenirs.
- Agents: Gemini, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/bgrossmankahn_facilitation-experiencedesign-ai-activity-7378510930810343424-LEfU (2025-09-29)
- Last verified: 2026-10-17

### Prepare a property tax appeal script and evidence packets with ChatGPT
- URL: https://www.realagentcases.com/use-cases/prepare-a-property-tax-appeal-script-and-evidence-packets-with-chatgpt/
- Outcome: A homeowner appealing a house assessed at $481,500 says he used ChatGPT to organise a 3-page hearing script and build three evidence packets (for himself, the Board and the Assessor) from comparable sales, repair quotes and listings. After presenting about 40 pages of evidence, he settled at $422,000, which he says lowers his bill by about $300–$500/year; the whole effort took about 6 hours.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ItsMikePoarch/status/1971680108982555068 (2025-09-26)
- Last verified: 2026-10-17

### Fix a car's dead AC and check-engine light with a code read and a blown fuse, guided by ChatGPT
- URL: https://www.realagentcases.com/use-cases/fix-a-car-s-dead-ac-and-check-engine-light-with-a-code-read-and-a-blown-fuse-gui/
- Outcome: When a check-engine light came on and the AC stopped working, Eric Pacifici (@SMB_Attorney) asked ChatGPT, which told them to get the code read at AutoZone. ChatGPT interpreted the code and walked them through replacing a blown fuse, which fixed it.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/SMB_Attorney/status/1967729428252987778 (2025-09-15)
- Last verified: 2026-10-17

### Get a $1,400 medical bill reduced with a note drafted by ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-1-400-medical-bill-reduced-with-a-chatgpt-drafted-note-citing-insurance-ru/
- Outcome: Adam Fern asked ChatGPT, 'Knowing what you know about insurance rules, how can I get this bill lower.' It produced a note, which Fern copied and pasted to the doctor, and says the bill went down by $1,400.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/fern/status/1966170040782504004 (2025-09-11)
- Last verified: 2026-10-17

### Test six AI homework tutors on one maths problem with a 14-year-old and pick ChatGPT Study Mode
- URL: https://www.realagentcases.com/use-cases/test-six-ai-homework-tutors-on-one-maths-problem-with-a-14-year-old-and-pick-cha/
- Outcome: During back-to-school prep, Laura Cioffi and their 14-year-old daughter gave the same cone surface-area problem to ChatGPT Study Mode, Photomath, Socratic by Google, Claude, Brainly AI and Duolingo Math. Photomath got it wrong; Claude and Socratic were right but too academic; Brainly gave clear steps; Duolingo felt like a game. ChatGPT Study Mode asked her for the values, walked her through the formula and checked her understanding; she said it didn't make her feel 'thick'.
- Agents: ChatGPT, Claude, Photomath, Socratic, Brainly AI, Duolingo Math
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test. A single-problem test by a parent, not a formal review.
- Source: https://www.linkedin.com/posts/lauracioffi_ai-education-parenting-activity-7371576560027582464-eqWm (2025-09-10)
- Last verified: 2026-10-17

### Plan gluten-free, braces-friendly family dinners and calendar invites with a Gemini Gem
- URL: https://www.realagentcases.com/use-cases/plan-gluten-free-braces-friendly-family-dinners-and-calendar-invites-with-a-gemi/
- Outcome: A parent set up a custom Gemini Gem as a family 'chef' that knows their needs (gluten free, braces, under 30 minutes, few dishes, in-season produce, macros). It builds the meal plan and shopping list and creates daily calendar invites with the recipe that sync to their family calendar. She calls it the most revolutionary AI use for her, though the lunch suggestions weren't a hit and still need refining.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/katrina-purcell_whats-for-dinner-the-most-dreaded-question-activity-7370800987680436226-jfSO (2025-09-08)
- Last verified: 2026-10-17

### Get a day-by-day moving plan with low-effort meals from ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-day-by-day-moving-plan-with-low-effort-meals-from-chatgpt/
- Outcome: Moving house, starting a new job and turning 40 within two weeks, a poster asked ChatGPT to act as her moving assistant and create a day-by-day plan up to the move. She shared her prompt, which asks for 3 to 5 priority tasks a day, a packing focus, admin tasks, 15-minute meals using what's left in the kitchen, a moving-day food plan and a first-day survival box checklist. She says the clear to-do list made everything feel more manageable.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Written before the move; the result isn't reported.
- Source: https://www.linkedin.com/posts/cianaabdollahian_too-many-life-events-and-not-enough-brain-activity-7365386039235866628-ZfV3 (2025-08-24)
- Last verified: 2026-10-17

### Make personalised storybooks and a party game for a baby shower with Gemini
- URL: https://www.realagentcases.com/use-cases/make-personalised-storybooks-and-a-party-game-for-a-baby-shower-with-gemini/
- Outcome: Organising a baby shower for three colleagues, a poster used Gemini's Storybook feature, with a photo and family names, to create a custom children's book for each family; they say the storylines were touching and the illustrations gorgeous. They also had Gemini write Family Feud-style survey questions and used NotebookLM for a comedy tribute video and an AI game-show co-host.
- Agents: Gemini, NotebookLM
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/frankieho18_how-i-threw-an-ai-powered-baby-shower-ive-activity-7365390880205070336-HD01 (2025-08-24)
- Last verified: 2026-10-17

### Turn a week of school emails, PDFs and WhatsApp messages into a one-page parent action plan with a ChatGPT prompt
- URL: https://www.realagentcases.com/use-cases/turn-a-week-of-school-emails-pdfs-and-whatsapp-messages-into-a-one-page-parent-a/
- Outcome: Claire Calvert shares a setup for parents buried in school communications: either a custom GPT called 'School PA' or a regular chat where you upload emails saved as PDFs, screenshots and forwarded WhatsApp messages each week. The prompt asks ChatGPT to extract every event, highlight parent decisions, tag deadlines by urgency with days remaining, and produce a conflict-free one-page weekly plan, without assuming information that isn't there.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task. The example schedule in the post is illustrative. The poster says they're building a product in this area.
- Source: https://www.linkedin.com/posts/clairefletcher2_turn-chatgpt-5-into-the-worlds-best-school-activity-7364629210717401088-Mye5 (2025-08-22)
- Last verified: 2026-10-17

### Plan a week of shellfish-free, macro-friendly meals with overlapping ingredients using Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-week-of-shellfish-free-macro-friendly-meals-with-overlapping-ingredients/
- Outcome: A poster asked Grok 4 for a week of meals for someone allergic to shellfish, hitting fitness macros, with 30 minutes max to cook and overlapping ingredients. The poster says that within 10 minutes they had 5 breakfasts, lunches and dinners with portions, one grocery list and a Sunday prep schedule, and after week one reports about 30% less grocery spending and 5+ hours saved, with less produce going to waste.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/caseywhalen14_5-hours-saved-30-less-grocery-spending-activity-7363932852083687424-otLz (2025-08-20)
- Last verified: 2026-10-17

### Diagnose a washing machine fault and get exact part numbers from Grok instead of calling a repairer
- URL: https://www.realagentcases.com/use-cases/diagnose-a-washing-machine-fault-and-get-exact-part-numbers-from-grok-instead-of/
- Outcome: Instead of calling a repairer or buying a new washing machine, the poster (@notswlfty1) asked Grok for help. Grok gave step-by-step diagnosis instructions and the exact part numbers needed, and the poster fixed the machine, estimating the saving at about $2,000.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The $2,000 saving is the poster's own estimate versus a repair visit or new machine.
- Source: https://x.com/notswlfty1/status/1957768166181920856 (2025-08-19)
- Last verified: 2026-10-17

### Keep two sick cats' vet updates, records and decisions straight in a ChatGPT project while away
- URL: https://www.realagentcases.com/use-cases/keep-two-sick-cats-vet-updates-records-and-decisions-straight-in-a-chatgpt-proje/
- Outcome: On holiday when both cats fell seriously ill, with their partner handling vet visits at home, Harriet Meyer set up a 'cat project' in ChatGPT. They uploaded the cats' medical histories, transcripts from vet calls and notes on temperament, and used it to make sense of updates, work out what to ask the vets, and think through difficult decisions. Meyer says it didn't replace the vets but helped them think clearly; one cat recovered, while the other was still in hospital when they posted.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/harriet-meyer_sometimes-the-most-unexpected-uses-for-ai-activity-7359467417737011200-aHQl (2025-08-08)
- Last verified: 2026-10-17

### Return a faulty used car for a full refund under UK consumer law with ChatGPT's help
- URL: https://www.realagentcases.com/use-cases/return-a-faulty-used-car-for-a-full-refund-under-uk-consumer-law-with-chatgpt-s/
- Outcome: Four weeks after buying a used car that kept developing problems, even after a warranty repair, a UK buyer thought he was stuck: the dealer's 30-day return only covered up to 250 miles and he'd done about 500. ChatGPT told him about his rights under the Consumer Rights Act 2015, and he says one email later the dealer agreed to take the car back with a full refund.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mike-garde_chatgpt-saved-me-from-keeping-a-car-i-no-activity-7356588894382911488-Px92 (2025-07-31)
- Last verified: 2026-10-17

### Get a Chevy Volt fault diagnosed on the roadside with Grok, before the dealer agreed
- URL: https://www.realagentcases.com/use-cases/get-a-chevy-volt-fault-diagnosed-on-the-roadside-with-grok-before-the-dealer-agr/
- Outcome: When a 2018 Chevy Volt had a problem, James Benson described it to Grok on the spot. Grok said the battery energy control module (BECM) needed replacing, that it was under extended warranty, and how to temporarily get home. The dealership initially said Grok was wrong, but after diagnosing the car found it was the BECM and covered by warranty.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Ichkabible/status/1951063168165519767 (2025-07-31)
- Last verified: 2026-10-17

### Shortlist a mid-size SUV and compare insurance policy clauses with NotebookLM
- URL: https://www.realagentcases.com/use-cases/shortlist-a-mid-size-suv-and-compare-insurance-policy-clauses-with-notebooklm/
- Outcome: Hunting for a mid-size SUV in India, Android Police's Parth Shah put long reviews and hours-long YouTube comparisons into a NotebookLM notebook and asked questions such as rear-seat comfort, service costs and A/C performance, with links back to each source. After settling on a car, Shah uploaded insurers' policy brochures and asked about accidental-damage exclusions, cancellation fees, cashless claims at service centers in Surat and natural-disaster cover, and says NotebookLM pulled out the exact clauses.
- Agents: NotebookLM
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/notebooklm-helped-me-finalize-car-purchase/ (2025-07-29)
- Last verified: 2026-10-17

### Stop overdrafting by staggering bills across paychecks with a ChatGPT budget plan
- URL: https://www.realagentcases.com/use-cases/stop-overdrafting-by-staggering-bills-across-paychecks-with-a-chatgpt-budget-pla/
- Outcome: A freelance writer who kept overdrafting despite a steady job listed her fixed expenses and due dates and asked ChatGPT for a monthly budget. It suggested splitting each paycheck, paying current bills with one part and saving the rest for bills due later, then broke down her remaining money into savings, debt repayment and personal spending. She says the first month was tough, but since then she hasn't overdrafted at all; she found she had to be very specific with her prompts.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/used-chatgpt-to-manage-finances-spend-better-save-money-2025-7 (2025-07-28)
- Last verified: 2026-10-17

### Book family movie tickets with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/book-family-movie-tickets-with-chatgpt-agent/
- Outcome: A poster says one prompt to ChatGPT agent booked family tickets for Fantastic Four that Sunday in a few steps. The cinema site's search didn't surface the movie correctly, but the agent found its way; it handed control back for him to fill in personal details before purchase, and took longer than a person would.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/vivek-bhatia-46057480_i-just-booked-movie-tickets-with-the-chatgpt-activity-7354391920774656000-Jmxx (2025-07-25)
- Last verified: 2026-10-17

### Book movie tickets, calendar events and a babysitter slot for date night with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/book-movie-tickets-calendar-events-and-a-babysitter-slot-for-date-night-with-cha/
- Outcome: TechRadar's Eric Hal Schwartz asked ChatGPT agent to plan a Saturday movie date: get tickets at a theater he named, add calendar events for dropping their child at her grandparents' and the showtime, and make an invitation for his wife. It asked which showing and seats he wanted, confirmed before buying, picked two good seats and paused at payment for him; then it created the calendar event, sent invites to his wife and her parents, and offered a separate babysitting event.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.techradar.com/computing/artificial-intelligence/i-tried-using-chatgpt-agent-to-plan-a-date-night-and-it-worked-surprisingly-well (2025-07-18)
- Last verified: 2026-10-17

### Fix a tankless water heater a company called 'shot' with Grok and ChatGPT
- URL: https://www.realagentcases.com/use-cases/fix-a-tankless-water-heater-a-company-called-shot-with-grok-and-chatgpt/
- Outcome: After a repair company said their tankless water heater was shot and quoted $6,900 to install a new one, a homeowner watched YouTube videos and chatted with Grok 4 and OpenAI's o3 model. The homeowner says they fixed it themselves the next morning in less than 30 minutes and then called the company to get their $100 back.
- Agents: Grok, ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/XTownSquareX/status/1945202539730968577 (2025-07-15)
- Last verified: 2026-10-17

### Plan a DIY 120-guest outdoor wedding with ChatGPT: to-do list, vendors, budget
- URL: https://www.realagentcases.com/use-cases/plan-a-diy-120-guest-outdoor-wedding-with-chatgpt-to-do-list-vendors-budget/
- Outcome: A journalist's write-up describes a bride planning a DIY, 120-person outdoor wedding in Las Vegas who says about 85 percent of the wedding to-dos were created, organised or completed with AI. Her request to ChatGPT for common and uncommon tasks produced a spreadsheet of 200 suggestions; it also helped with the day-of schedule, seating chart, menu, website text, vendor query letters and budget, and found the officiant and cake maker. She says it sometimes suggested vendors no longer in town, which a phone call cleared up.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A journalist's post written before the wedding; the 85% figure is the bride's estimate.
- Source: https://www.linkedin.com/posts/alix-strauss-52842a1b_how-ai-is-transforming-wedding-planning-activity-7350140327258648576-SGE9 (2025-07-13)
- Last verified: 2026-10-17

### Plan a 20th anniversary party, from crowd-size recipes to a helper playbook, with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-20th-anniversary-party-from-crowd-size-recipes-to-a-helper-playbook-with/
- Outcome: Hosting a big 20th wedding anniversary party, a poster used ChatGPT to build the event plan, name signature cocktails and mocktails, convert a pan-fried crab cake recipe into mini air-fried versions for a crowd, create a gluten-free dessert a guest couldn't tell was gluten-free, write a playbook for the teenagers hired to help, write signage and checklists, plan the evening timeline with a vow renewal, and even fix a printer jam.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/njshelsby_we-threw-a-big-party-for-our-20th-wedding-activity-7348510547866198017-LjlR (2025-07-09)
- Last verified: 2026-10-17

### Win a $2,100 settlement from a tow driver who dented a car, with guidance from ChatGPT and Gemini
- URL: https://www.realagentcases.com/use-cases/win-a-2-100-settlement-from-a-tow-driver-who-dented-a-car-with-guidance-from-cha/
- Outcome: With no pre-tow photo and a dent left by a tow truck driver hired by his apartment, a poster asked ChatGPT o3 and Gemini 2.5 Flash whether to claim on insurance or leave it. Both suggested settling directly with the driver, who also lacked a pre-tow photo. Over five weeks, with o3's guidance, he built an email trail, set deadlines and asked for evidence, and says he won a $2100 settlement in 35 days.
- Agents: ChatGPT, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/williameerdmans_genai-chatgpt-o3-activity-7348421812793262080-FX2c (2025-07-08)
- Last verified: 2026-10-17

### Catch up on a long group-trip email thread with Gemini in Gmail
- URL: https://www.realagentcases.com/use-cases/catch-up-on-a-long-group-trip-email-thread-with-gemini-in-gmail/
- Outcome: Looped into a never-ending email thread about planning a weekend trip, Android Police's Ben Khalesi tapped Gemini's summary in Gmail instead of reading about 20 messages. He says it returned a neat recap: the Airbnb in Perth booked for July 12–14, who was handling groceries, the 9 am meeting point and a reminder to bring sunscreen.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/how-gemini-fixed-my-inbox/ (2025-07-04)
- Last verified: 2026-10-17

### Find out a broken AC is just a bad thermostat with Grok and Gemini
- URL: https://www.realagentcases.com/use-cases/find-out-a-broken-ac-is-just-a-bad-thermostat-with-grok-and-gemini/
- Outcome: When their AC stopped working and none of half a dozen companies could come before Friday, a poster asked Grok and Gemini. The poster says they correctly diagnosed a broken thermostat, probably saving hundreds of dollars and 2 days without AC.
- Agents: Grok, Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AZinCLE/status/1937946028340109341 (2025-06-25)
- Last verified: 2026-10-17

### Pick a new country to move to from a list of must-haves with ChatGPT
- URL: https://www.realagentcases.com/use-cases/pick-a-new-country-to-move-to-from-a-list-of-must-haves-with-chatgpt/
- Outcome: Needing to leave Canada, a poster and his partner gave ChatGPT a list of requirements: what they loved about Vancouver, economic viability, quality of life and more. It suggested a handful of options with Munich at the top, and they moved there with little German, having never spent more than 24 hours in the city.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written on arrival; how the move went isn't reported.
- Source: https://www.linkedin.com/posts/swilsonnz_i-moved-country-because-of-chatgpt-seriously-activity-7340445564037156865-f6rV (2025-06-16)
- Last verified: 2026-10-17

### Turn news stories into bedtime summaries a 9-year-old understands with Gemini
- URL: https://www.realagentcases.com/use-cases/turn-news-stories-into-bedtime-summaries-a-9-year-old-understands-with-gemini/
- Outcome: A parent in Canada pastes news links into Gemini with prompts that turn them into kid-friendly summaries, short stories with relatable characters, or bullet-point takeaways for his nine-year-old son. He says a wildfire story came out relatable without scary imagery, and a screen-time story helped his son understand why too much gaming can lead to frustration.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-use-gemini-to-explain-the-world-to-my-kids-these-3-prompts-nail-it-every-time (2025-06-15)
- Last verified: 2026-10-17

### Get out of a $75 parking ticket with a polite ChatGPT-written appeal
- URL: https://www.realagentcases.com/use-cases/get-out-of-a-75-parking-ticket-with-a-polite-chatgpt-written-appeal/
- Outcome: A parent asked ChatGPT to write a polite appeal explaining that his children distracted him and he genuinely forgot (which he says was true). He says it got it in one try, he sent it through the appeal form, and woke up to an email saying his $75 balance had been forgiven.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/davidsawyer1/status/1932183768334258353 (2025-06-09)
- Last verified: 2026-10-17

### Compare ChatGPT and Gemini for healthy meal ideas and a week of cheap lunches
- URL: https://www.realagentcases.com/use-cases/compare-chatgpt-and-gemini-for-healthy-meal-ideas-and-a-week-of-cheap-lunches/
- Outcome: Tom's Guide's Josh Render photographed chicken thighs, salad and avocados and asked ChatGPT and Gemini for a healthy meal with those ingredients, then asked each for a week of varied, healthy, low-cost lunches for losing weight. He preferred Gemini for both: it explained how to prepare the chicken beforehand and was more comprehensive, though wordier; ChatGPT's terser plan was easier to print and hang on the fridge.
- Agents: Gemini, ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/i-put-chatgpt-vs-google-gemini-to-the-test-to-see-which-is-better-at-meal-prep-heres-the-winner (2025-05-30)
- Last verified: 2026-10-17

### Build a colour-coded exam revision timetable for a 12-year-old with Gemini
- URL: https://www.realagentcases.com/use-cases/build-a-colour-coded-exam-revision-timetable-for-a-12-year-old-with-gemini/
- Outcome: A parent and his 12-year-old daughter gathered her subjects, revision guides and exam timetable, including photos of crumpled papers from her school bag, and gave Gemini 2.5 Pro detailed instructions on subject priorities and daily study time. He says it produced a detailed, accurate plan that became a colour-coded web page she can view anywhere, marking exam days differently without being asked; it took about 15 minutes and left her more enthusiastic about revising.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/fergusadam_love-it-when-you-push-ai-to-do-something-activity-7323249464435978240-97Xb (2025-04-30)
- Last verified: 2026-10-17

### Set up tap-to-see box inventories with NFC tags for a move with ChatGPT
- URL: https://www.realagentcases.com/use-cases/set-up-tap-to-see-box-inventories-with-nfc-tags-for-a-move-with-chatgpt/
- Outcome: Moving out with belongings stored at a friend's house while staying with family before three weeks in Europe, Nicole Leffer asked ChatGPT for a way to know what was in each box without opening it. It suggested NFC sticker tags (100 for about $26) and walked through setting up iPhone Shortcuts so each tag opens a numbered note. As boxes were packed, including by movers, each note got the box's destination, photos and a list of contents. Tapping a box now shows what's inside, and searching notes found an external monitor without opening boxes.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/nicoleleffer_i-moved-on-wednesday-and-chatgpt-has-made-activity-7321683668660334594-Di_l (2025-04-25)
- Last verified: 2026-10-17

### Organise symptoms, past tests and questions before a specialist appointment with ChatGPT
- URL: https://www.realagentcases.com/use-cases/organise-symptoms-past-tests-and-questions-before-a-specialist-appointment-with/
- Outcome: Before a gastroenterologist appointment, a patient used ChatGPT to map out her symptoms, track the tests she'd already had and write the specific questions she wanted to ask, so she wouldn't forget anything in a short visit. Afterwards, it helped her turn the doctor's guidance into a personal care plan she could follow.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/meganmarsiglio_i-had-an-appointment-with-my-gastroenterologist-activity-7320534051704946688-D-n6 (2025-04-22)
- Last verified: 2026-10-17

### Get an $800 refund for unwanted post-trial streaming charges with a ChatGPT-drafted legal notice
- URL: https://www.realagentcases.com/use-cases/get-an-800-refund-for-unwanted-post-trial-streaming-charges-with-a-chatgpt-draft/
- Outcome: After a video streaming service kept charging him for months once a free trial ended, without sending receipts or payment alerts, an India-based poster says ChatGPT helped him draft a legal notice and he got an $800 refund.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/641947jeet/status/1912657499775246696 (2025-04-17)
- Last verified: 2026-10-17

### Decode 100+ pages of new-apartment purchase paperwork and plan the layout and lighting with Gemini and NotebookLM
- URL: https://www.realagentcases.com/use-cases/decode-100-pages-of-new-apartment-purchase-paperwork-and-plan-the-layout-and-lig/
- Outcome: Moving into a newly bought apartment, Mitja Rutnik uploaded the purchase documents to Gemini to ask about confusing clauses, such as a warranty policy that only referenced laws, and set up a NotebookLM notebook with the full 100+ page document set for quick look-ups. Gemini also advised on colour schemes, furniture proportions, counter-to-table clearance, rug size and lumens and colour temperature per room, suggested how to tackle bathroom floor stains, summarized warranty rights when raising defects with the developer, and gave layout ideas from the floor plan. Rutnik still had to build a 3D model in a separate app.
- Agents: Gemini, NotebookLM
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidauthority.com/i-moved-gemini-help-3543960/ (2025-04-17)
- Last verified: 2026-10-17

### Build a streaming bundle under $30 a month that keeps Netflix with ChatGPT
- URL: https://www.realagentcases.com/use-cases/build-a-streaming-bundle-under-30-a-month-that-keeps-netflix-with-chatgpt/
- Outcome: Tom's Guide streaming writer Alix Blackburn asked ChatGPT for a budget-friendly streaming bundle under $30 that always includes Netflix. It suggested keeping Netflix constant and rotating others, pointed to the Hulu and Disney Plus bundle, offered a monthly rotation plan she hadn't asked for and included free services. She notes it forgot she reviews streaming for work, suggested overly minimal bundles, and mentioned Crackle, which has shut down.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tomsguide.com/entertainment/streaming/i-asked-chatgpt-to-help-me-save-money-on-streaming-heres-what-happened (2025-04-16)
- Last verified: 2026-10-17

### Manage a parent's Parkinson's and dementia care paperwork and family updates with ChatGPT
- URL: https://www.realagentcases.com/use-cases/manage-a-parent-s-parkinson-s-and-dementia-care-paperwork-and-family-updates-wit/
- Outcome: When their mother was diagnosed with Parkinson's and dementia, Matt Strain and their brother shared the caregiving. Strain used ChatGPT from diagnosis to their mother's passing to simplify medical information, clarify medications and contraindications, streamline insurance paperwork, and communicate with medical teams and family; Strain wrote an accompanying article with the prompts used.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mattstrain_how-ai-became-my-unexpected-ally-in-caregiving-activity-7318299136208818178-OAeb (2025-04-16)
- Last verified: 2026-10-17

### Narrow a truck's no-start down to a dead battery with ChatGPT
- URL: https://www.realagentcases.com/use-cases/narrow-a-truck-s-no-start-down-to-a-dead-battery-with-chatgpt/
- Outcome: When his F250 wouldn't start (lights on, no click, no check-engine light), a poster described the symptoms to ChatGPT and within 15 minutes it narrowed things to the starter motor or the battery. AutoZone's free test showed the battery was dead; he swapped it and the truck started, saving about $1,000 in shop or towing fees by his estimate.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ATX_UG_GUY/status/1911976084091904165 (2025-04-15)
- Last verified: 2026-10-17

### Work out which federal tax forms to file and how to fill them, line by line, with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-out-which-federal-tax-forms-to-file-and-how-to-fill-them-line-by-line-with/
- Outcome: A filer gave ChatGPT a rundown of his W2, 1099 and business income, expenses and stock gains and losses, and asked which forms and schedules to use, in what order, with line-by-line instructions. He says it helped a lot with forms, sequence and deductions such as the self-employment tax deduction, but badly mixed up stock gains and losses numbers; he caught it, and ChatGPT admitted and fixed the mistake.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Helpful overall, but it swapped stock gain/loss numbers until he caught it.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/lafrancegreg_i-used-chatgpt-to-do-my-federal-taxes-initially-activity-7317775164706840576-JD7g (2025-04-15)
- Last verified: 2026-10-17

### Turn a spring-cleaning checklist into a doable 1-hour-a-day plan with ChatGPT
- URL: https://www.realagentcases.com/use-cases/turn-a-spring-cleaning-checklist-into-a-doable-1-hour-a-day-plan-with-chatgpt/
- Outcome: Tom's Guide's Kaycee Hill asked ChatGPT for a complete spring cleaning checklist for a one-bedroom apartment and got room-by-room sections including easy-to-miss tasks. She then asked it to split the plan into 5 days with 1 hour of cleaning per day; it saved heavier jobs like deep-cleaning the bathroom for the weekend and put quick wins on weekdays. She says she followed the plan and, for once, finished.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-used-chatgpt-tasks-to-create-the-perfect-spring-cleaning-schedule-heres-how (2025-04-11)
- Last verified: 2026-10-17

### Sort a $4,288 dealer repair estimate into needed work and likely upsells with ChatGPT
- URL: https://www.realagentcases.com/use-cases/sort-a-4-288-dealer-repair-estimate-into-needed-work-and-likely-upsells-with-cha/
- Outcome: Cord Silverstein uploaded a dealer's $4,288 estimate for a 2019 Honda Pilot, including a vague photo labelled 'leaking tensioner', and asked ChatGPT to act as a mechanic. It said the coolant flush was legitimate, to ask for photo proof of the timing belt and tensioner leak, that the shock absorbers looked like an upsell because the photo showed a new part, and that alignment was only needed if the shocks were bad. It also explained Honda's service intervals and drafted a polite pushback email.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The '$3K+' saving in the post is ChatGPT's estimate; the post doesn't say what work was finally done or paid for.
- Source: https://www.linkedin.com/posts/cordsilverstein_when-the-car-dealership-handed-my-wife-a-activity-7316601466838745088-hdiT (2025-04-11)
- Last verified: 2026-10-17

### Use ChatGPT as a 'case manager' across three hospitals for a parent's care: jargon, questions, pain tracking and handover notes
- URL: https://www.realagentcases.com/use-cases/use-chatgpt-as-a-case-manager-across-three-hospitals-for-a-parent-s-care-jargon/
- Outcome: During three months of a mother's illness across three hospitals, Dror Grof fed ChatGPT charts, doctors' notes, medications and daily updates. After realizing medical decisions belonged with the care team, Grof used it to explain jargon, suggest what to ask next, set up a subjective pain scale (some staff asked to see it as a graph), keep one source of truth when doctors changed, transcribe some conversations, and draft transition reports between facilities.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster also relied on a senior doctor as an adviser and says trust for medical decisions belongs with the medical team. Check consent rules before recording any conversation with medical staff.
- Source: https://www.linkedin.com/posts/dror-grof_digitalhealth-ai-activity-7314995369501728768-FSj8 (2025-04-07)
- Last verified: 2026-10-17

### Get one misclassified blood-panel charge reversed on a live insurer call with ChatGPT's help
- URL: https://www.realagentcases.com/use-cases/get-one-misclassified-blood-panel-charge-reversed-on-a-live-insurer-call-with-ch/
- Outcome: Billed for a standard blood panel from a routine preventive visit that had been misclassified, Kevin Darbelnet kept ChatGPT open during a call with UnitedHealthcare. It decoded the diagnosis and CPT codes, fed Darbelnet questions to ask the provider and insurer in real time, drafted a formal appeal letter and helped assemble EOBs, provider statements and claim history. One charge was reversed during the call; an appeal on the second was submitted within an hour and was still pending at the time of posting.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. One of the two charges was reversed during the call; the appeal on the second was still pending.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster is an AI strategy consultant, and the post ends with a pitch for that work.
- Source: https://www.linkedin.com/posts/kevin-darbelnet_aistrategy-chatgpt-innovationinaction-activity-7313581208343326720-5z-o (2025-04-03)
- Last verified: 2026-10-17

### Choose an insurance-covered breast pump and nursery basics with Grok
- URL: https://www.realagentcases.com/use-cases/choose-an-insurance-covered-breast-pump-and-nursery-basics-with-grok/
- Outcome: A first-time mom whose insurance offered 2 pages of free or heavily discounted breast pumps had no idea where to start. She says Grok narrowed her choices by asking a few good questions, suggested a $5 fridge thermometer so milk stored in her wonky dorm fridge stays safe, and recommended a lamp and baby monitor.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/MelissasRPGs/status/1905325791069548633 (2025-03-27)
- Last verified: 2026-10-17

### Find appointments and add emailed invites to your calendar with Gemini in Gmail
- URL: https://www.realagentcases.com/use-cases/find-appointments-and-add-emailed-invites-to-your-calendar-with-gemini-in-gmail/
- Outcome: Tom's Guide's Christoph Schwaiger walks through using Gemini in Gmail (a paid feature at the time) to declutter an inbox, including asking 'Search through my emails, when is my next dental appointment?' to fish out forgotten appointment details, and 'Create an event in my calendar based on this email' for invitations that arrive as emails rather than calendar invites. It's a how-to guide, not a report of results.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.tomsguide.com/ai/i-used-google-gemini-to-declutter-my-gmail-account-heres-how-you-can-do-it-too (2025-03-24)
- Last verified: 2026-10-17

### Get a faulty car sound system replaced free after a $10K dealer quote, with a ChatGPT complaint email
- URL: https://www.realagentcases.com/use-cases/get-a-faulty-car-sound-system-replaced-free-after-a-10k-dealer-quote-with-a-chat/
- Outcome: Out of warranty and quoted $10K by Subaru to replace a faulty sound system, a poster says he fed ChatGPT Australian consumer law guidance and the warranty terms, asked it to find related complaints and tribunal cases, and had it draft a firm email. He says 48 hours after sending it, Subaru called and offered to replace it at no charge.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Written in a promotional tone; not independently confirmed.
- Source: https://www.linkedin.com/posts/mrdavediamond_subaru-quoted-me-10k-chatgpt-fixed-it-in-activity-7309527361803165697-yJRl (2025-03-23)
- Last verified: 2026-10-17

### Answer destination-wedding guests' questions with a custom wedding GPT in ChatGPT
- URL: https://www.realagentcases.com/use-cases/answer-destination-wedding-guests-questions-with-a-custom-wedding-gpt-in-chatgpt/
- Outcome: Planning a destination wedding in Sesimbra, Portugal, and getting the same questions from guests despite a detailed website, a bride built a custom GPT as a travel assistant: she pasted in the private wedding website content and told it to act as a friendly travel agent answering about venues, travel tips and dress code. She says it took under 5 minutes to set up and saves her time; custom GPTs need a paid account.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/sarahurbonas_weddingplanning-chatgpt-activity-7306348398167695360-ridw (2025-03-14)
- Last verified: 2026-10-17

### Repair a chewed garden hose with Grok's measuring tips
- URL: https://www.realagentcases.com/use-cases/repair-a-chewed-garden-hose-with-grok-s-measuring-tips/
- Outcome: After squirrels chewed through their garden hose, a poster described the problem to Grok. The poster says it explained how to measure the hose's inside diameter and what the standard hose-fitting repair looks like, and the fix was a piece of cake.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AmpegBabyBass76/status/1899609109046972505 (2025-03-11)
- Last verified: 2026-10-17

### Build a university decision matrix for a daughter choosing between four schools, using Grok for data
- URL: https://www.realagentcases.com/use-cases/build-a-university-decision-matrix-for-a-daughter-choosing-between-four-schools/
- Outcome: A parent built a one-page decision matrix to help his daughter choose between the four universities that accepted her, using Grok as the data source and factoring in Florida's Bright Futures scholarships. He shared it so she could adjust the scoring and add criteria herself.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/markzolna_my-daughters-university-choice-decision-activity-7303423775834845184-3x7x (2025-03-06)
- Last verified: 2026-10-17

### Trace a car transmission problem to a faulty fuse with Grok
- URL: https://www.realagentcases.com/use-cases/trace-a-car-transmission-problem-to-a-faulty-fuse-with-grok/
- Outcome: A driver with a transmission problem talked it through with Grok, which suggested troubleshooting steps. The driver says the cause turned out to be a faulty fuse, avoiding an expensive repair shop bill.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/4evrshepherds/status/1896656129733882315 (2025-03-03)
- Last verified: 2026-10-17

### Narrow down an intermittent furnace problem before the HVAC visit with Grok
- URL: https://www.realagentcases.com/use-cases/narrow-down-an-intermittent-furnace-problem-before-the-hvac-visit-with-grok/
- Outcome: With her furnace acting up intermittently again, a poster ran all the symptoms and everything she'd tried through Grok. She says it walked her through the most likely issues and narrowed them down a lot; even if she couldn't fix it herself, it armed her with information to cut troubleshooting time so the HVAC company would come prepared.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/RareCamellia/status/1892447755265556869 (2025-02-20)
- Last verified: 2026-10-17

### Check a kindergartener's math worksheets and make harder practice problems with ChatGPT
- URL: https://www.realagentcases.com/use-cases/check-a-kindergartener-s-math-worksheets-and-make-harder-practice-problems-with/
- Outcome: On a snow day, a parent uploaded his five-year-old's two completed math worksheets to ChatGPT, asking it to act as his kindergarten teacher, check the answers, give feedback and create more challenging problems if needed. ChatGPT marked all the answers correct and suggested moving on to addition within 20, missing-number problems and two-step word problems.
- Agents: ChatGPT
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/jeremiah-kalir-phd_second-remote-learning-day-for-my-five-year-activity-7298330729359323140--aAh (2025-02-20)
- Last verified: 2026-10-17

### Track a dishwasher leak to a busted drain hose with ChatGPT (a pro finished the job)
- URL: https://www.realagentcases.com/use-cases/track-a-dishwasher-leak-to-a-busted-drain-hose-with-chatgpt-a-pro-finished-the-j/
- Outcome: When a dishwasher leak came through his basement ceiling, a homeowner shared photos with ChatGPT and, over a few back-and-forths, narrowed it to a busted drain hose and a worn insulation piece. It found the right parts for his model with a purchase link and walked him through the fix. On the next cycle another puddle appeared because he hadn't clamped the hose far enough, and he called a professional; he says AI got him '99% there'.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Diagnosis and part were right, but the hose wasn't clamped far enough and he called a pro to finish.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/artembeer_ai-homerepair-practicaltech-activity-7298022017348378625-UQse (2025-02-19)
- Last verified: 2026-10-17

### Plan a new rescue puppy's supplies, potty schedule and training with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-new-rescue-puppy-s-supplies-potty-schedule-and-training-with-chatgpt/
- Outcome: After spontaneously adopting a puppy, a poster says ChatGPT listed everything they needed at the pet supply shop in seconds, and since then they've used it for night-time toilet schedules based on her last meal, training tips for basic commands, ways to encourage eating, and techniques to calm her crying.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/benjamin-briggs-4b4828101_yes-im-shamelessly-using-my-new-puppy-activity-7295037223811338242-WakW (2025-02-11)
- Last verified: 2026-10-17

### Generate IEP meeting questions and measurable goals from your child's current IEP with a ChatGPT prompt
- URL: https://www.realagentcases.com/use-cases/generate-iep-meeting-questions-and-measurable-goals-from-your-child-s-current-ie/
- Outcome: Raquel Contreras-Vides, a parent of a neurodivergent son, shares a ChatGPT prompt for annual IEP (ARD) meetings: ChatGPT takes the role of a special education advocate in your state, reviews the attached IEP, goals and accommodations, and generates key questions and quantitative goals, asking you for details if it needs them.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task. The post's claim that goals will be 'legally defensible' is the poster's view; check with a qualified advocate.
- Source: https://www.linkedin.com/posts/raquelcontrerasvides_neurospicyadvocate-iep-specialeducation-activity-7295122533689831424-S3PI (2025-02-11)
- Last verified: 2026-10-17

### Plan a Super Bowl party menu, room layout and halftime games: ChatGPT vs DeepSeek
- URL: https://www.realagentcases.com/use-cases/plan-a-super-bowl-party-menu-room-layout-and-halftime-games-chatgpt-vs-deepseek/
- Outcome: A Tom's Guide editor gave ChatGPT and DeepSeek three prompts for a 15-person Super Bowl party: menu with quantities and prep timeline, living-room layout, and halftime games. She found ChatGPT strongest on precise portions, timelines and organised lists (U-shaped seating, TV at 42-48 inches, commercial bingo), while DeepSeek offered more creative ideas and better dietary alternatives and traffic-flow tips, though some suggestions like three charcuterie boards felt like overkill.
- Agents: ChatGPT, DeepSeek
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-and-deepseek-to-plan-my-super-bowl-party-heres-what-happened (2025-02-01)
- Last verified: 2026-10-17

### Cut 23,000 unread emails to about 8,000 with one Gemini command in Gmail
- URL: https://www.realagentcases.com/use-cases/cut-23-000-unread-emails-to-about-8-000-with-one-gemini-command-in-gmail/
- Outcome: Faced with more than 23,000 unread emails, David Alonge asked Gemini in Gmail to 'Delete all unread emails from 2009 to 2022.' Within a few minutes the unread count dropped to about 8,000, a job Alonge expected would have taken weeks by hand.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The poster didn't say whether anything was reviewed before deletion; the stronger version adds a check first.
- Source: https://www.linkedin.com/posts/davidalonge_what-could-have-taken-me-2-weeks-ai-did-activity-7290638847292379137-acgb (2025-01-30)
- Last verified: 2026-10-17

### Recreate grandma's bread pudding from memory with Grok
- URL: https://www.realagentcases.com/use-cases/recreate-grandma-s-bread-pudding-from-memory-with-grok/
- Outcome: Years after last making their grandma's bread pudding, a poster gave Grok the basics of the recipe they remembered instead of searching for hours. The poster says the recipe Grok came back with was exactly what they were looking for.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/WaskelweeWabbit/status/1870558445599895823 (2024-12-21)
- Last verified: 2026-10-17

### Win a small-claims case against a landlord over a deposit with ChatGPT's help
- URL: https://www.realagentcases.com/use-cases/win-a-small-claims-case-against-a-landlord-over-a-deposit-with-chatgpt-s-help/
- Outcome: A tenant says ChatGPT flagged that their landlord's security deposit exceeded legal limits under a 2019 law. When the landlord refused to return the money and filed a retaliatory $5,000 counterclaim, the tenant used ChatGPT to build the case, represented themselves in small claims court and says they won $1,180, with the counterclaim dismissed. The poster adds that AI tools are not substitutes for legal advice.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/MatchbookSeries/status/1864355869502394479 (2024-12-04)
- Last verified: 2026-10-17

### Log new board game purchases for BoardGameGeek with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/log-new-board-game-purchases-and-add-them-to-boardgamegeek-with-claude-in-chrome/
- Outcome: From one instruction, the agent found the BoardGameGeek IDs of the three ordered games, called Claude in Chrome twice and wrote a board game note (with title and description) for each; the screenshot doesn't show the games being added to the BoardGameGeek account.
- Agents: Claude in Chrome
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/dSebastien/status/2110255458934301089 (2026-10-14)
- Last verified: 2026-10-16

### Sort a week of unread email into reply, money and clutter with Gemini
- URL: https://www.realagentcases.com/use-cases/sort-a-week-of-unread-email-into-reply-money-and-clutter-with-gemini/
- Outcome: Android Police's Akshay Bhalla, facing thousands of unread emails, set up a Gemini scheduled action that summarizes the last seven days of unread Gmail into three groups: threads needing his reply, financial emails like invoices and receipts, and low-priority newsletters and alerts. He says the near-daily report puts the few emails needing a reply front and center; with a second prompt to purge clutter, he says he cleaned 200 unread threads in an hour.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Article or interview
- Claimed result: 200 unread threads cleaned in one hour (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/gemini-gmail-prompt-clean-email-threads/ (2026-10-14)
- Last verified: 2026-10-16

### Check whether a noisy dishwasher is failing from a recording with Claude
- URL: https://www.realagentcases.com/use-cases/check-whether-a-noisy-dishwasher-is-failing-from-a-recording-with-claude/
- Outcome: A Hacker News commenter noticed a loud dishwasher at an old apartment and, unsure whether to call a technician, recorded the noise, screenshotted a spectrogram (a picture of the sound) from an app, took photos and gave them to Claude. The commenter says it matched frequencies to the machine's components and concluded it was operating normally, with the pump just louder from age and wear. A reply asked how they could know Claude was right without verifying.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: A reply on the same thread questioned how the commenter could verify Claude's conclusion; the commenter answered that they have general expertise in adjacent fields to evaluate it and that being wrong was low-risk. Treat the verdict as a guide and keep watching the machine.
- Source: https://news.ycombinator.com/item?id=50057799 (2026-10-12)
- Last verified: 2026-10-16

### Phone a local organization with a question using Genspark
- URL: https://www.realagentcases.com/use-cases/phone-a-local-organization-with-a-question-using-genspark/
- Outcome: In a filmed test, Genspark's calling agent phoned a church, asked whether it had a young adults group and what ages it served, got a clear answer (Friday nights, ages 18–31) and a follow-up about another group, then returned a transcript and summary; a separate restaurant booking call stalled.
- Agents: Genspark
- Category: Personal
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://missiondrivenbrand.com/ai-phone-agent-for-outbound-calls/ (2026-10-08)
- Last verified: 2026-10-16

### Write the next gym training cycle from workout logs with Claude Code
- URL: https://www.realagentcases.com/use-cases/write-the-next-gym-training-cycle-from-workout-logs-with-claude-code/
- Outcome: Claude Code read four weeks of the poster's workout logs, found where lifts stalled, wrote the next training cycle with a reason for each starting weight, loaded the routines into the tracking app, and noted where the app would suggest the wrong weight.
- Agents: Claude Code
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Moretti_ale90/status/2106692994942308605 (2026-10-04)
- Last verified: 2026-10-16

### Get a Tesla's 12V power outlet working again with Grok
- URL: https://www.realagentcases.com/use-cases/get-a-tesla-s-12v-power-outlet-working-again-with-grok/
- Outcome: An X user says they wanted to plug in a Starlink but think they fried something by plugging 12V gear into the Tesla's 16V cigarette-lighter outlet, and didn't expect to get a service appointment in time. They say Grok troubleshot the problem and helped them reset the eFuse, and the Tesla is fixed.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/profs77/status/2106886795027702237 (2026-10-04)
- Last verified: 2026-10-16

### Rename a pile of train tickets by date and destination with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/rename-a-pile-of-train-tickets-by-date-and-destination-with-claude-cowork/
- Outcome: Claude Cowork read a pile of train tickets and renamed each file by its correct date and destination.
- Agents: Claude Cowork
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/airocket2026/status/2104541855815991507 (2026-09-28)
- Last verified: 2026-10-16

### Turn an inbox into a daily to-do list with Claude Cowork (abandoned)
- URL: https://www.realagentcases.com/use-cases/turn-an-inbox-into-a-daily-to-do-list-with-claude-cowork-abandoned/
- Outcome: Claude Cowork summarized the poster's email into a to-do list, but the list grew long and unmanageable, so the poster stopped reading it and turned it off as a waste of tokens.
- Agents: Claude Cowork
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The to-do list grew too long to use, so the poster stopped reading it and turned it off.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/uliang6482/status/2101965199452610582 (2026-09-21)
- Last verified: 2026-10-16

### Find extra apartment charges in email and request a refund with Grok Bot
- URL: https://www.realagentcases.com/use-cases/find-extra-apartment-charges-in-email-and-request-a-refund-with-grok-bot/
- Outcome: The poster says that while getting him a parking spot, Grok Bot found about $986 of extraneous charges from his apartment in his email and emailed the building asking for a refund; he shared a screenshot.
- Agents: Grok Bot
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/Baconbrix/status/2101058998527279242 (2026-09-18)
- Last verified: 2026-10-16

### Shop for cheaper home insurance quotes with Claude Code and Chrome
- URL: https://www.realagentcases.com/use-cases/shop-for-cheaper-home-insurance-quotes-with-claude-code-and-chrome/
- Outcome: Using the poster's coverage details and access to Chrome, Gmail and the existing policy, Claude Code sought quotes from about a dozen insurers and found two strong candidates; the poster says they were approved by AAA at $3,100/year, a $4,600/year saving.
- Agents: Claude Code, Claude in Chrome
- Category: Personal
- Evidence: Described by poster
- Claimed result: $4,600/year saved (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/sethprattsf/status/2100023090860429659 (2026-09-16)
- Last verified: 2026-10-16

### Use Gemini Live's camera to troubleshoot a modem or appliance
- URL: https://www.realagentcases.com/use-cases/use-gemini-live-s-camera-to-troubleshoot-a-modem-or-appliance/
- Outcome: Google's guide to household chores with Gemini suggests pointing Gemini Live's camera at a problem, for example asking 'What does this blinking light on the modem mean?', to get step-by-step guidance on setting up toys, caring for a car or repairing appliances. It also suggests scanning fridge shelves for leftover-based meal ideas and uploading receipt photos to track spending. It offers prompts only and doesn't report results.
- Agents: Gemini Live
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://blog.google/innovation-and-ai/products/gemini-app/household-chores-tips/ (2026-09-15)
- Last verified: 2026-10-16

### Troubleshoot a misbehaving espresso machine with Grok
- URL: https://www.realagentcases.com/use-cases/troubleshoot-a-misbehaving-espresso-machine-with-grok/
- Outcome: An X user gives Grok a shout-out for helping her fix her HP printer and then, the 'biggest challenge of all', her Breville espresso machine, which she says now works better than it has in years.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/TXRealtorDallas/status/2095979978995450326 (2026-09-04)
- Last verified: 2026-10-16

### File a utility complaint about low voltage with Claude
- URL: https://www.realagentcases.com/use-cases/file-a-utility-complaint-about-low-voltage-with-claude/
- Outcome: A Hacker News commenter who had circuit-level power monitoring set up at home was analyzing bills when he mentioned voltage problems to Claude. He says it put together a packet documenting brownouts down to 95v and routine dips below 110v, and helped him fill out a Board of Public Utilities complaint form. Two days later the utility, JCP&L, called; he says he still needs to call back to ask for a larger transformer.
- Agents: Claude
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49535164 (2026-09-02)
- Last verified: 2026-10-16

### Set up a voice-mode study tutor for young kids with ChatGPT
- URL: https://www.realagentcases.com/use-cases/set-up-a-voice-mode-study-tutor-for-young-kids-with-chatgpt/
- Outcome: Chris Domanski, father of two daughters aged 8 and 6, photographs their study guides and has ChatGPT's voice mode quiz them for weekly math and spelling tests, varying questions, revisiting misses and explaining concepts different ways, with guardrails to keep it on task. He says through trial and error it saved time and he thinks it improved their grades.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/dad-chatgpt-tutor-helps-my-kids-get-better-grades-2026-8 (2026-08-24)
- Last verified: 2026-10-16

### Keep a move on track with a running ChatGPT to-do list
- URL: https://www.realagentcases.com/use-cases/keep-a-move-on-track-with-a-running-chatgpt-to-do-list/
- Outcome: A Japanese note.com writer used ChatGPT to manage the many small, time-sensitive tasks of a move, starting with 'Summarize what I need to do for my move' and adding things as they came up, like using up the fridge. She says she did the procedures herself but barely had to think about what to do next.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://note.com/maple_yamazaki/n/n731be2c56b67?hl=en (2026-08-15)
- Last verified: 2026-10-16

### Screen and clear an email inbox in one pass with Claude Code
- URL: https://www.realagentcases.com/use-cases/screen-and-clear-an-email-inbox-in-one-pass-with-claude-code/
- Outcome: The poster runs all email in Claude Code: in one pass it screened 27 emails, archived and unsubscribed, and sent sub-agents after the three that needed real thought.
- Agents: Claude Code
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/francoisdefitte/status/2086831437047566793 (2026-08-10)
- Last verified: 2026-10-16

### Organize a Spanish visa renewal and fill the forms with Claude and Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/organize-a-spanish-visa-renewal-and-fill-the-forms-with-claude-and-claude-in-chr/
- Outcome: For a Spanish visa renewal, the poster set up a bureaucracy folder with Claude, fed in the family's information, had Claude verify documents, then used Claude in Chrome to fill the online forms faster and more accurately than checking by hand.
- Agents: Claude, Claude in Chrome
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/milanm_/status/2071962695264288973 (2026-06-30)
- Last verified: 2026-10-16

### Get a furnace running until the repair visit with Gemini and a video
- URL: https://www.realagentcases.com/use-cases/get-a-furnace-running-until-the-repair-visit-with-gemini-and-a-video/
- Outcome: With no repair appointment available for 2 days over the 2025 holidays, a Hacker News commenter filmed the furnace trying to start and gave the videos to Gemini. The commenter says it diagnosed the issue immediately and had them spin a small exhaust fan while the furnace tried to fire; it came on, and repeating that kept it working until the HVAC service arrived.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The prompt is quoted from the commenter's follow-up reply on the same Hacker News thread.
- Source: https://news.ycombinator.com/item?id=48417845 (2026-06-05)
- Last verified: 2026-10-16

### Diagnose an air conditioner that stopped cooling with ChatGPT and photos
- URL: https://www.realagentcases.com/use-cases/diagnose-an-air-conditioner-that-stopped-cooling-with-chatgpt-and-photos/
- Outcome: A landlord whose tenant reported the AC wasn't cooling couldn't get a callback from his HVAC company, so he photographed the outdoor compressor panel and used ChatGPT. He says it walked him through what to check using the diagnostic codes he saw and through replacing a $25 capacitor; the service call alone would have cost almost 4x that.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: $25 part vs a service call almost 4x that (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48419105 (2026-06-05)
- Last verified: 2026-10-16

### Fix a gas dryer that won't heat with Gemini and a photo
- URL: https://www.realagentcases.com/use-cases/fix-a-gas-dryer-that-won-t-heat-with-gemini-and-a-photo/
- Outcome: A Hacker News commenter's gas dryer ran but wouldn't heat. Gemini suggested a thermal fuse, pointed it out in a photo of the opened back panel, walked him through testing it, linked the part once he gave the model number, and advised cleaning the vent since poor venting likely blew the fuse. He says after cleaning the exhaust and a $5 fuse the dryer works fine.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: $5 fuse (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=48419007 (2026-06-05)
- Last verified: 2026-10-16

### Generate a homeschool week's assignments from state standards with Gemini
- URL: https://www.realagentcases.com/use-cases/generate-a-homeschool-week-s-assignments-from-state-standards-with-gemini/
- Outcome: Kyle Elliott used Gemini with the state's reference material to draft a week-by-week lesson plan for a 15-year-old daughter, then adjusted it by hand for learning differences, weeks off and her 504 (school accommodations) plan, and saved it. Each week an automation Elliott built sends that week's topics, curriculum standards, homeschool laws and her interests to Gemini, which writes the full assignment list around what she cares about.
- Agents: Gemini
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://blog.kyleelliott.net/the-ai-homeschool-principal-how-i-built-a-system-that-plans-my-daughters-entire-school-week/ (2026-05-07)
- Last verified: 2026-10-16

### Transcribe and organize 411 family voice memos with Claude Code
- URL: https://www.realagentcases.com/use-cases/transcribe-and-organize-411-family-voice-memos-with-claude-code/
- Outcome: Claude Code transcribed and clustered 411 voice memos left by the poster's late grandmother and built a small site organized by mood and date, which the family uses more than the photo album.
- Agents: Claude Code
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: 411 voice memos transcribed (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/the_smart_ape/status/2051926703287533820 (2026-05-06)
- Last verified: 2026-10-16

### Prepare questions for a doctor's visit and a travel packing list with Grok
- URL: https://www.realagentcases.com/use-cases/prepare-questions-for-a-doctor-s-visit-and-a-travel-packing-list-with-grok/
- Outcome: An X user lists what Grok helped her with over a couple of days: looking at over-the-counter medication options and drawing up questions for her next doctor's visit, planning purchases and a packing list for a cross-country trip to a funeral, and planning her next batch of poetry submissions.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AnnDiegoUSA/status/2047829048987152808 (2026-04-25)
- Last verified: 2026-10-16

### Have ChatGPT review a finished tax return for missed savings
- URL: https://www.realagentcases.com/use-cases/have-chatgpt-review-a-finished-tax-return-for-missed-savings/
- Outcome: Matt Koppenheffer redacted his TurboTax return and worksheets and asked ChatGPT in a temporary chat whether they were under-optimized or should take other actions. The first 17-minute run found they had taken a foreign tax deduction instead of a foreign tax credit; switching raised state tax but cut federal tax more, saving money. A couple more hours of back-and-forth produced no further savings, but he says it raised his confidence the return was correct.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://mattkoppenheffer.substack.com/p/i-put-my-tax-return-into-chatgpt (2026-04-13)
- Last verified: 2026-10-16

### Appeal a surprise dental bill with ChatGPT
- URL: https://www.realagentcases.com/use-cases/appeal-a-surprise-dental-bill-with-chatgpt/
- Outcome: After a new dentist's visit left her owing $1,201.80 on top of the $2,111.29 she'd already paid, Jamie Phillis gave ChatGPT her call transcripts, receipt, explanation of benefits and her insurer's appeal instructions. It wrote a formatted appeal letter in under a minute and flagged duplicate evaluations and a charge at seven times the allowed rate. The first appeal was denied; ChatGPT then drafted a second-level appeal and a complaint to the Arizona Department of Insurance. She says the case was never formally resolved, but the office never sent her a bill for the uncovered balance and she never paid the $1,200.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: $1,201.80 balance never billed after appeals (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/used-ai-help-me-dispute-medical-bill-1200-2026-3 (2026-03-30)
- Last verified: 2026-10-16

### Plan Saturday errands and find haircut slots with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/plan-saturday-errands-and-find-haircut-slots-with-chatgpt-agent/
- Outcome: Asked to plan Saturday errands (a haircut, an Amazon return, groceries and Target), ChatGPT grouped stops by location into an efficient route, and the agent found nearby kid-friendly salons, opened booking pages and narrowed to the earliest slots; the author notes it can't always finalize bookings.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-agent-to-plan-my-weekend-errands-and-it-even-booked-a-haircut-for-me (2026-03-07)
- Last verified: 2026-10-16

### Build a price matrix to cut a pet insurance premium with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/build-a-price-matrix-to-cut-a-pet-insurance-premium-with-claude-in-chrome/
- Outcome: After Embrace raised the poster's premium to $549 a month for her dog, Claude in Chrome navigated the insurer's site, tested every coverage combination, built a price matrix and recommended cancelling; she says that saved about $6,500 a year.
- Agents: Claude in Chrome
- Category: Personal
- Evidence: Shown on video
- Claimed result: About $6,500 a year saved, per the poster (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/natalieevagray/status/2028956353851773017 (2026-03-03)
- Last verified: 2026-10-16

### Book a doctor's appointment with agent mode (blocked)
- URL: https://www.realagentcases.com/use-cases/book-a-doctor-s-appointment-with-agent-mode-blocked/
- Outcome: The poster tried to set up a doctor's appointment with agent mode and got nowhere; they say ChatGPT and Manus agents are held back by blocking and captchas.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Got nowhere; the poster says blocking and captchas stopped it.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/NanjuanS/status/2025030709887512754 (2026-02-21)
- Last verified: 2026-10-16

### Draft a firm landlord notice about a repeat water leak with Claude Code
- URL: https://www.realagentcases.com/use-cases/draft-a-firm-landlord-notice-about-a-repeat-water-leak-with-claude-code/
- Outcome: Instead of paying a lawyer for a strongly worded notice, the poster gave Claude Code the lease documents and context about a second water leak; it drafted an email that got them the concessions they wanted.
- Agents: Claude Code
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/4484/status/2024645227089428607 (2026-02-20)
- Last verified: 2026-10-16

### Find a Texas electricity plan on Power to Choose with Gemini in Chrome
- URL: https://www.realagentcases.com/use-cases/find-a-texas-electricity-plan-on-power-to-choose-with-gemini-in-chrome/
- Outcome: Ars Technica had Chrome's Auto Browse search powertochoose.org for a 12–24 month plan for 2,000 kWh a month with a specific utility; it set the filters, sorted results and returned a recommended plan with its fact sheet in a couple of minutes. The reviewer scored it 10/10.
- Agents: Gemini in Chrome (auto browse)
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://arstechnica.com/google/2026/02/tested-how-chromes-auto-browse-agent-handles-common-web-tasks/ (2026-02-12)
- Last verified: 2026-10-16

### Choose and walk through a cheap tax e-file with Grok
- URL: https://www.realagentcases.com/use-cases/choose-and-walk-through-a-cheap-tax-e-file-with-grok/
- Outcome: An X user says she sent Grok one message about filing federal and state taxes for a basic household (salaried husband, part-time independent contractor), asking for an efficient, cheap e-file with direct deposit. She says Grok guided her step by step, she filed for under $50, and she saved $200+ compared with past years.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/AHebert57196/status/2021042080881967458 (2026-02-10)
- Last verified: 2026-10-16

### Turn a dinner idea into a Google Keep grocery checklist with Gemini
- URL: https://www.realagentcases.com/use-cases/turn-a-dinner-idea-into-a-google-keep-grocery-checklist-with-gemini/
- Outcome: MakeUseOf's Saikat Basu asked Gemini, with the Keep connector, for a 'week of high-protein vegetarian dinners grocery list' for one person. Gemini created the checklist in Google Keep item by item; he says it was neatly organized, though it didn't group items by produce and pantry within the note.
- Agents: Gemini
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.makeuseof.com/gemini-makes-google-keep-more-useful/ (2026-01-09)
- Last verified: 2026-10-16

### Find a cheaper part for a $2,800 suspension repair quote with ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-a-cheaper-part-for-a-2-800-suspension-repair-quote-with-chatgpt/
- Outcome: USA Today reports that TikTok user everything_bylaneise was quoted $2,800 by a mechanic for a suspension issue, and ChatGPT told her she could source the part from a junkyard for $50. An electrician by trade with the right tools, she documented doing the labor herself and, by the article's count, saved $2,750. The article cautions that chatbots can miss steps and this won't work for everyone.
- Agents: ChatGPT
- Category: Personal
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.usatoday.com/story/cars/maintenance/2025/12/06/chatgpt-diy-car-repair-save-money/87485355007/ (2025-12-06)
- Last verified: 2026-10-16

### Make a printable family calendar and kids' chore chart with ChatGPT
- URL: https://www.realagentcases.com/use-cases/make-a-printable-family-calendar-and-kids-chore-chart-with-chatgpt/
- Outcome: Bitdefender's parenting guide shares ready-to-copy ChatGPT prompts for everyday parent tasks, including turning a list of activities into a printable Monday–Sunday family calendar with colour labels and making a fun weekly chore chart for a 10-year-old. It offers prompts only and doesn't report results.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.bitdefender.com/en-us/blog/hotforsecurity/7-everyday-parent-tasks-chatgpt-can-simplify-with-prompts-to-help-you-do-it (2025-12-04)
- Last verified: 2026-10-16

### Audit a bank statement for zombie subscriptions and price creep with ChatGPT
- URL: https://www.realagentcases.com/use-cases/audit-a-bank-statement-for-zombie-subscriptions-and-price-creep-with-chatgpt/
- Outcome: A banker writing for Evem Finance fed a sanitized bank statement to ChatGPT with a 'forensic accountant' prompt looking for repeating subscriptions, bills that crept up over 6 months, and duplicate charges or unusual fees. The author says it found $200 a month in wasted spending, or $2,400 a year, in 30 seconds.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: $200 a month ($2,400 a year) found (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://evem.org/i-used-chatgpt-to-find-2400-in-hidden-waste-in-my-bank-statement-here-is-the-exact-prompt/ (2025-11-23)
- Last verified: 2026-10-16

### Plan seven quick, healthy dinners for two with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-seven-quick-healthy-dinners-for-two-with-chatgpt/
- Outcome: CNET's tester asked ChatGPT for a seven-day plan of quick, healthy dinners for two people and a grocery list. All seven dinners could be cooked in under 45 minutes; ingredients came without measurements, and she tweaked seasoning and cooking order on some recipes. She says grocery shopping became easier and the couple wasted no food that week.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.cnet.com/home/kitchen-and-household/how-i-used-chatgpt-to-craft-a-budget-friendly-meal-plan/ (2025-11-22)
- Last verified: 2026-10-16

### Work through an insurance external appeal for a child's surgery with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-through-an-insurance-external-appeal-for-a-child-s-surgery-with-chatgpt/
- Outcome: A Hacker News commenter says that after insurance declined a life-saving surgery for their 6-year-old at the last minute, ChatGPT guided them through the external appeal process, suggested who to contact to apply pressure, researched questions and helped with wording. They didn't follow every suggestion and checked steps with friends and family; the procedure was approved 10 days after the denial.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=45736945 (2025-10-28)
- Last verified: 2026-10-16

### Plan a cheap, healthy week of meals for two with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-cheap-healthy-week-of-meals-for-two-with-chatgpt/
- Outcome: Tom's Guide's Alex Hughes asked ChatGPT for a Monday-to-Friday plan for two in the UK, shopping at Sainsbury's, cheap, healthy and filling. After re-prompting for more interesting recipes, then less focus on calories and cost, he says it nailed the brief, with recipes and one ordered shopping list marking optional extras, though some recipes were a bit vague until refined. The whole process took about 10-15 minutes.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-to-plan-my-meals-for-a-week-it-saved-me-money-and-hours-of-hassle (2025-10-27)
- Last verified: 2026-10-16

### Apply to LinkedIn jobs from a resume with Perplexity Comet
- URL: https://www.realagentcases.com/use-cases/apply-to-linkedin-jobs-from-a-resume-with-perplexity-comet/
- Outcome: Comparing Comet and ChatGPT Atlas for LinkedIn job applications, the poster found Comet far better; with a detailed prompt and their resume, it applied to 20+ jobs and sent notes to HR.
- Agents: Perplexity Comet
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: 20+ job applications (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/vedanthpeace/status/1982463132980392390 (2025-10-26)
- Last verified: 2026-10-16

### Get tips for budgeting holiday spending with Grok
- URL: https://www.realagentcases.com/use-cases/get-tips-for-budgeting-holiday-spending-with-grok/
- Outcome: GOBankingRates asked Grok for its best tips on budgeting for holiday expenses. Grok suggested setting a total covering gifts, travel, food and decorations, with a common guideline of 1% to 2% of annual income, saving a small monthly amount ahead of time, tracking spending with an app like YNAB or a spreadsheet, and cutting gift spending in favour of travel or hosting if family time matters most.
- Agents: Grok
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.gobankingrates.com/saving-money/holiday/asked-grok-how-to-budget-for-holiday-expenses-heres-what-it-said/ (2025-10-19)
- Last verified: 2026-10-16

### Request a LinkedIn Premium refund with Comet
- URL: https://www.realagentcases.com/use-cases/request-a-linkedin-premium-refund-with-comet/
- Outcome: Asked to find LinkedIn's refund page and issue a refund for a just-purchased Premium subscription, Comet instead emailed RevenueCat for a refund without the poster approving that final action.
- Agents: Perplexity Comet Assistant
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Sent a refund email to the wrong company (RevenueCat) without asking first.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/andreimatei/status/1979395146413412686 (2025-10-18)
- Last verified: 2026-10-16

### Spot when you avoid conflict in meeting recordings with Claude Code
- URL: https://www.realagentcases.com/use-cases/spot-when-you-avoid-conflict-in-meeting-recordings-with-claude-code/
- Outcome: Dan Shipper puts all his meeting recordings in a folder and asks Claude Code to find every time he subtly avoided conflict.
- Agents: Claude Code
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.lennysnewsletter.com/p/everyone-should-be-using-claude-code (2025-10-14)
- Last verified: 2026-10-16

### Diagnose why your Mac is running slow with Claude Code
- URL: https://www.realagentcases.com/use-cases/diagnose-why-your-mac-is-running-slow-with-claude-code/
- Outcome: Anthony Roux uses Claude Code to check load, memory pressure, disk space, stuck processes and swap, dig into causes like caches, Docker and Time Machine snapshots, explain the results and suggest fixes with risk notes.
- Agents: Claude Code
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.lennysnewsletter.com/p/everyone-should-be-using-claude-code (2025-10-14)
- Last verified: 2026-10-16

### Investigate a wet spot on a wall with Gemini Live's camera
- URL: https://www.realagentcases.com/use-cases/investigate-a-wet-spot-on-a-wall-with-gemini-live-s-camera/
- Outcome: A new homeowner writing for All Things Open showed Gemini a wet spot on his bathroom wall through a live camera feed, explained the context and answered its clarifying questions. Considering nearby windows and the roof, Gemini walked him through a structured assessment and recommended calling a roofer to inspect a possible roof leak.
- Agents: Gemini Live
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://allthingsopen.org/articles/ai-troubleshoot-household-problems (2025-10-10)
- Last verified: 2026-10-16

### Cancel Netflix, Disney+ and Peacock with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/cancel-netflix-disney-and-peacock-with-chatgpt-agent/
- Outcome: Tom's Guide's Amanda Caswell used ChatGPT agent to cancel three subscriptions—Peacock, Disney and Netflix—entering her credentials at the login screens herself. She says she had no problems cancelling, but warns that some sites have anti-bot defenses or don't recognize the agent, and advises double-checking before it confirms.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tomsguide.com/ai/i-used-chatgpt-to-cancel-my-streaming-subscriptions-heres-how-to-try-it-too (2025-10-03)
- Last verified: 2026-10-16

### Organize 12,000 photos by date and find duplicates with Claude Code
- URL: https://www.realagentcases.com/use-cases/organize-12-000-photos-by-date-and-find-duplicates-with-claude-code/
- Outcome: Asked to organize 12,000+ photos by EXIF date and location, find duplicates and group burst shots, Claude Code built year/month folders and an event summary, found 3,200 duplicates and 890 burst shots, and surfaced photos the author thought were lost, in about twenty minutes.
- Agents: Claude Code
- Category: Personal
- Evidence: Article or interview
- Claimed result: 3,200 duplicates found (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://newsletter.artofsaience.com/p/claude-code-is-not-a-coding-toolits (2025-09-20)
- Last verified: 2026-10-16

### Find who to escalate a stuck insurance bill to with ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-who-to-escalate-a-stuck-insurance-bill-to-with-chatgpt/
- Outcome: Medical Xpress reports that Lauren Gonzalves spent almost two years fighting an $1,800 Anthem bill for her newborn's hospital nursery care. After a free AI appeal assistant (Counterforce's Maxwell) pointed to supporting reasons, she asked ChatGPT how to escalate a grievance with Anthem and got email addresses at a company handling the insurer's complaints. Within days an analyst found the claim, and a week later the bill was dismissed, she says.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://medicalxpress.com/news/2025-09-company-denied-medical-ai-victory.html (2025-09-17)
- Last verified: 2026-10-16

### Estimate how much to save for retirement in 20 years with Grok
- URL: https://www.realagentcases.com/use-cases/estimate-how-much-to-save-for-retirement-in-20-years-with-grok/
- Outcome: GOBankingRates asked Grok about a hypothetical 47-year-old earning $72,000 with no retirement savings, retiring at 67. Grok estimated $866,575 would be needed, assuming 70% of income in retirement ($91,028 a year in 2045 dollars at 3% inflation), an estimated Social Security benefit of $56,365 (2024 dollars), a $34,663 annual gap and a 4% withdrawal rate.
- Agents: Grok
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Caution: this is a hypothetical profile set by GOBankingRates, not a real person's plan. Every figure, including the $56,365 Social Security benefit, is Grok's estimate and wasn't independently checked; the article suggests looking up your own estimate in your My Social Security account.
- Source: https://www.gobankingrates.com/retirement/planning/asked-grok-how-much-money-needed-retire-20-years/ (2025-09-12)
- Last verified: 2026-10-16

### Notify Germany's TV licence (GEZ) office of a house move with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/notify-germany-s-tv-licence-office-of-a-house-move-with-chatgpt-agent/
- Outcome: After a few problems, ChatGPT agent found the GEZ (public broadcasting fee) contact form, wrote a message and entered the new address; the poster only had to solve one captcha.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kimmonismus/status/1961927860538515829 (2025-08-30)
- Last verified: 2026-10-16

### Design a 7-day high-protein, low-carb meal plan with Grok
- URL: https://www.realagentcases.com/use-cases/design-a-7-day-high-protein-low-carb-meal-plan-with-grok/
- Outcome: An X user says she used Grok to design a 7-day meal plan around her goals, sharing her constraints: 100 grams of protein a day, low carb, low-lectin vegetables, goat cheese, A2 dairy, healthy fats, two meals a day, fermented foods and berries. She calls it an amazing tool.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/eveforamerica/status/1958118756644802825 (2025-08-20)
- Last verified: 2026-10-16

### Plan a taco-night grocery list by voice with Grok in a Tesla
- URL: https://www.realagentcases.com/use-cases/plan-a-taco-night-grocery-list-by-voice-with-grok-in-a-tesla/
- Outcome: Jake Lowney says that while driving his Model Y he asked Grok to help plan a grocery list for carne asada tacos, guacamole and chips for dinner with his kids. He says it nailed the list, added a local beer suggestion, and saved him mental bandwidth before the shopping trip.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/jake-lowney-csp-b64b9640_grok-tesla-xai-activity-7362528754255974401-53jo (2025-08-16)
- Last verified: 2026-10-16

### Find a restaurant-week steak deal and book a table for 2 with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-a-restaurant-week-steak-deal-and-book-a-table-for-2-with-chatgpt-agent/
- Outcome: Asked to find the best Dine LA steak dinner deal and book for 2 the next day, ChatGPT agent booked Dear John's for 8:15, though it misread the menu by treating sides as a separate course.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/wittinglyai/status/1953481169707610268 (2025-08-07)
- Last verified: 2026-10-16

### Pay a UK car tax reminder with ChatGPT agent (handoff for payment)
- URL: https://www.realagentcases.com/use-cases/pay-a-uk-car-tax-reminder-with-chatgpt-agent-handoff-for-payment/
- Outcome: From a photo of the posted reminder and a chosen payment option, ChatGPT agent found the government website and filled out the forms; the poster took over to enter card details.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/pauldm/status/1950157691084517393 (2025-07-29)
- Last verified: 2026-10-16

### Apply to a job with ChatGPT agent (small-button trouble)
- URL: https://www.realagentcases.com/use-cases/apply-to-a-job-with-chatgpt-agent-small-button-trouble/
- Outcome: Testing ChatGPT agent on a real job application (to OpenAI), the poster found it much better than the earlier Operator but it struggled to click small buttons; they rated Manus as slightly ahead at times.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Struggled to click small buttons during the application.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/affaan/status/1949007302193029325 (2025-07-26)
- Last verified: 2026-10-16

### Write a firm but friendly repair-request letter to a landlord with ChatGPT
- URL: https://www.realagentcases.com/use-cases/write-a-firm-but-friendly-repair-request-letter-to-a-landlord-with-chatgpt/
- Outcome: Writer Deborah Sosin says her rented apartment's dishwasher broke in December 2024 and stayed broken despite multiple polite requests. With a new lease pending she asked ChatGPT to help write a forceful letter; it framed her priorities as keeping the tenancy, protecting her health, and getting the dishwasher replaced without triggering hostility. In a later update she says she had a cordial in-person talk with the landlady, a brand-new dishwasher was on the premises to be installed, and 'the letter worked.'
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://deborahs27.substack.com/p/how-a-bot-helped-me-write-a-disgruntled (2025-07-25)
- Last verified: 2026-10-16

### Troubleshoot a scooter that won't start with a Grok checklist
- URL: https://www.realagentcases.com/use-cases/troubleshoot-a-scooter-that-won-t-start-with-a-grok-checklist/
- Outcome: Todd Robbins asked Grok for a troubleshooting matrix for a friend's 2006 Vespa GT200 that wouldn't start. Testing showed no fuel and no spark; the matrix's immobilizer column suggested trying a spare key, and when his friend fetched one the scooter started right away. He says a web search for the same matrix gave scattered, cursory results.
- Agents: Grok
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://themanonthemove.substack.com/p/grok-fixes-the-vespa (2025-07-23)
- Last verified: 2026-10-16

### Set up a puppy feeding, potty and medicine schedule with ChatGPT
- URL: https://www.realagentcases.com/use-cases/set-up-a-puppy-feeding-potty-and-medicine-schedule-with-chatgpt/
- Outcome: In his first week with 13-week-old Great Dane puppy Ruby, Tony of Tony Reviews Things used ChatGPT for crate setup, training 'sit' and a feeding schedule that included meds and potty breaks. When he described her symptoms it listed possibilities with a UTI at the top; he raised it with the vet, who confirmed a UTI, and she was back to normal within 24 hours of treatment. He says it gave him the language to advocate at the vet, not a replacement for one.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tonyreviewsthings.com/chatgpt-raise-puppy/ (2025-06-30)
- Last verified: 2026-10-16

### Tailor a résumé, cover letters and mock interviews for a job hunt with ChatGPT
- URL: https://www.realagentcases.com/use-cases/tailor-a-r-sum-cover-letters-and-mock-interviews-for-a-job-hunt-with-chatgpt/
- Outcome: Returning to job hunting after 10 years, Santiago Barraza Lopez used ChatGPT to pinpoint roles at the intersection of legal practice and emerging tech, rewrite his résumé against target job descriptions within minutes, draft tailored cover letters, and run mock interviews with role-specific questions and feedback. He says he eventually landed a role in AI.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/landed-dream-job-using-chatgpt-write-resume-cover-letter-2025-6 (2025-06-08)
- Last verified: 2026-10-16

### Draft a security-deposit demand letter to a silent landlord with ChatGPT
- URL: https://www.realagentcases.com/use-cases/draft-a-security-deposit-demand-letter-to-a-silent-landlord-with-chatgpt/
- Outcome: The Economic Times reports that a tenant whose landlord stopped responding after move-out used GPT-4 to draft a formal letter demanding the return of a Rs 35,000 security deposit, after lawyers wouldn't take such a small case. The AI asked for names, address, amount, length of stay, rent and past contact attempts, and wrote a letter citing consumer laws and warning of legal action. The tenant says the landlord paid within a few days, and that they cross-checked the laws themselves after others warned about made-up legal references.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://economictimes.indiatimes.com/magazines/panache/tenant-recovers-rs-35000-security-deposit-from-landlord-using-chatgpt-but-some-netizens-question-the-method/articleshow/121571382.cms (2025-06-02)
- Last verified: 2026-10-16

### Clean up a Gmail inbox with Gemini's search prompts (hit or miss)
- URL: https://www.realagentcases.com/use-cases/clean-up-a-gmail-inbox-with-gemini-s-search-prompts-hit-or-miss/
- Outcome: Tom's Guide's John Brandon found Gemini in Gmail accurate for composing emails and summarizing threads, but its inbox search and clean-up prompts were unreliable: it mostly searched recent mail, listed newsletters as his most-discussed topics, refused to find the largest attachments, and found only four emails with shipping labels out of several dozen. 'Show me all emails with an attachment from May 2024' did work, and he says clean-up worked about 25% of the time.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Search and clean-up prompts worked only about 25% of the time and mostly looked at recent emails.
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-am-using-gemini-now-for-my-gmail-and-theres-one-major-discovery-thats-surprising (2025-06-01)
- Last verified: 2026-10-16

### Plan a gluten-free week of Trader Joe's meals on $125 with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-gluten-free-week-of-trader-joe-s-meals-on-125-with-chatgpt/
- Outcome: Outside's Julia Ries Wexler asked ChatGPT for a week of gluten-free meals and snacks, plus dog treats, with a Trader Joe's grocery list, item names and prices on a $125 budget. She says the plan was missing items needed for the recipes, prices were inaccurate and two items weren't in stock; her receipt came to $160, about $35 over budget.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Missing ingredients, inaccurate prices and two out-of-stock items; the receipt was about $35 over budget.
- Prompt type: Original (as published in the source)
- Source: https://www.outsideonline.com/health/nutrition/chatgpt-meal-planning/ (2025-05-16)
- Last verified: 2026-10-16

### Find forgotten subscriptions on a credit card statement with ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-forgotten-subscriptions-on-a-credit-card-statement-with-chatgpt/
- Outcome: Brian Foster exported his latest statement, redacted sensitive details in macOS Preview, and asked ChatGPT to identify all monthly charges, sort them by service, amount and frequency, and suggest which to cancel. He says it flagged the recurring charges within seconds, grouped them so he could spot ones he didn't remember or hadn't used in months, and suggested which to cancel.
- Agents: ChatGPT
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://glassalmanac.com/i-let-ai-manage-my-bank-account-to-detect-useless-charges/ (2025-05-15)
- Last verified: 2026-10-16

### Plan five days of meals and a $75 grocery list with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-five-days-of-meals-and-a-75-grocery-list-with-chatgpt/
- Outcome: Delish's writer asked ChatGPT for Monday–Friday recipes (breakfast, lunch, dinner and snacks) and a grocery list on a $75 budget with no tomatoes, peas or oats, then asked it to drop tofu. The list came to about $74.50, but she says it had her buy five items she didn't need, and by Wednesday the chicken-heavy meals felt repetitive; she decided the method wasn't for her.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The list included five items she didn't need, and she found the meals repetitive.
- Prompt type: Original (as published in the source)
- Source: https://www.delish.com/food-news/a64236165/ai-chatgpt-meal-plan-grocery-list/ (2025-03-26)
- Last verified: 2026-10-16

### Get a weekly digest of school emails with Gemini in Gmail
- URL: https://www.realagentcases.com/use-cases/get-a-weekly-digest-of-school-emails-with-gemini-in-gmail/
- Outcome: TechRadar's writer stopped trying to organize his inbox and instead asks Gemini in Gmail targeted questions; one favourite is asking what he needs to know about his sons' school this week, which returns a curated list from the many mostly irrelevant school emails. He says Gemini in Gmail saves him about 5 hours a week.
- Agents: Gemini in Gmail
- Category: Personal
- Evidence: Described by poster
- Claimed result: About 5 hours a week saved (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://tech.yahoo.com/ai/articles/used-gemini-ai-declutter-gmail-163243622.html (2025-03-12)
- Last verified: 2026-10-16

### Find a mobile mechanic for a dead car battery with Grok
- URL: https://www.realagentcases.com/use-cases/find-a-mobile-mechanic-for-a-dead-car-battery-with-grok/
- Outcome: An X user says that at 8 am his car battery was dead; at 8:15 he asked Grok to find a mobile mechanic, at 8:17 he tapped the phone number, the mechanic arrived at 9 am and the battery was replaced by 9:30 am.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/brucevanzyl/status/1895309490272248091 (2025-02-28)
- Last verified: 2026-10-16

### Rewrite a resume from three reviewers' points of view with Grok
- URL: https://www.realagentcases.com/use-cases/rewrite-a-resume-from-three-reviewers-points-of-view-with-grok/
- Outcome: Andrew Donato uploaded his own draft resume so Grok 3 could match his voice, then asked it to critique the resume as a fast-screening in-house recruiter, an outsourced recruiter sourcing a senior sales leadership role, and a CRO. At each step he asked how likely that reader was to reach out and tweaked toward a 9/10, then had Grok merge the feedback in his original tone. He says the two-page result has already gotten him a few responses.
- Agents: Grok
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/andrewmdonato_been-really-enjoying-grok3-from-x-this-week-activity-7298729643572903937-AqiJ (2025-02-21)
- Last verified: 2026-10-16

### Make a complete moving-house checklist with Perplexity
- URL: https://www.realagentcases.com/use-cases/make-a-complete-moving-house-checklist-with-perplexity/
- Outcome: With a house closing the next day and a move a month away, the Get On With It blogger asked Perplexity for an exhaustive, categorized, copyable checklist covering admin, address changes and tasks they might not think of, like checking the new furnace filter. Perplexity returned a checkable list split into pre-move, moving-day and post-move tasks, with cited sources.
- Agents: Perplexity
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://getonwithit.blog/2025/02/18/a-checklist-for-moving-house-made-with-ai/ (2025-02-18)
- Last verified: 2026-10-16

### Work out how much to save per paycheck for a cross-country move with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-out-how-much-to-save-per-paycheck-for-a-cross-country-move-with-chatgpt/
- Outcome: Mashable's writer asked ChatGPT how much she needed to save from each paycheck to move over 1,100 miles to Chicago, with rent of $1,000-$1,500 in mind. She says the math was spotty—unless the prompt specified July 1, 2024 it assumed other dates, and once it calculated 10 pay periods instead of the 7 she gave—but after tweaking it gave a number that matched her own math. Apartment-search GPTs were a worse experience than Googling.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The math was spotty: it assumed wrong dates and once used 10 pay periods instead of 7 until she tweaked the prompt.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://me.mashable.com/tech/39912/i-tried-using-chatgpt-to-help-me-move-across-the-country (2024-03-26)
- Last verified: 2026-10-16

### Work out what check-engine codes mean and what to test, with ChatGPT
- URL: https://www.realagentcases.com/use-cases/work-out-what-check-engine-codes-mean-and-what-to-test-with-chatgpt/
- Outcome: SlashGear describes YouTuber Watch Wes Work asking ChatGPT about three codes (P0171, P0420, P1101) on a 2015 Chevrolet Trax. ChatGPT suggested checking for exhaust leaks, then inspecting the mass airflow sensor and air intake, and gave theoretically accurate test methods, but its normal airflow reading was vague. The YouTuber found vacuum leaks in a PCV hose and the valve cover and, asked whether ChatGPT diagnosed the issue, said it's a matter of perspective, but ultimately it did not; SlashGear says it lacked hands-on context for this model.
- Agents: ChatGPT
- Category: Personal
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT gave general tests but, by the YouTuber's account, did not diagnose the cause; its airflow figure was vague.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.slashgear.com/1391055/can-chat-gpt-help-fix-your-car/ (2023-09-16)
- Last verified: 2026-10-16

### Turn kids' activities into a daily family schedule with ChatGPT
- URL: https://www.realagentcases.com/use-cases/turn-kids-activities-into-a-daily-family-schedule-with-chatgpt/
- Outcome: Mom Ijeoma Nwaogu gives ChatGPT her children's routines, for example that they return from school at 4:30 p.m., have snacks before tutoring at 6 p.m., and it generates a detailed daily schedule; she also uses it for recipes for picky eaters from what's in the house and for assigning chores.
- Agents: ChatGPT
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/chatgpt-parenting-meal-planning-family-schedule-mom-guilt-2023-5 (2023-05-15)
- Last verified: 2026-10-16

### Track and cancel unused subscriptions with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/track-and-cancel-unused-subscriptions-with-chatgpt-agent/
- Outcome: ChatGPT agent scans for recurring charges and reminds the author to cancel before renewal, helping cancel unused streaming services.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/5-clever-ways-i-use-chatgpt-agent-to-save-serious-money-each-month (2025-10-10)
- Last verified: 2026-10-16

### Find the best flight options on Skyscanner with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-the-best-flight-options-on-skyscanner-with-chatgpt-agent/
- Outcome: ChatGPT agent searched Skyscanner across flexible dates and returned the flight the author would actually book in 3 of 3 tests, though airline sites and 3D Secure blocked it from completing the booking.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://wow.pjh.is/journal/can-chatgpt-agent-find-a-flight (2025-08-22)
- Last verified: 2026-10-16

### Find an available domain name with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-an-available-domain-name-with-chatgpt-agent/
- Outcome: The poster found ChatGPT agent useful for finding an available domain name that met a set of constraints.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/davlgd/status/1955070135833620532 (2025-08-12)
- Last verified: 2026-10-16

### Book a business-lunch restaurant reservation with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/book-a-business-lunch-restaurant-reservation-with-chatgpt-agent/
- Outcome: From a terse prompt, ChatGPT agent found a restaurant, filled in the form (including a shellfish allergy), and got a confirmed reservation in 11 minutes versus about 2 manually.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.nngroup.com/articles/impressions-chatgpt-agent/ (2025-08-08)
- Last verified: 2026-10-16

### Sign up for a Steam account with Comet Assistant
- URL: https://www.realagentcases.com/use-cases/sign-up-for-a-steam-account-with-comet-assistant/
- Outcome: Comet Assistant created a Steam account with the email the poster gave it, and solved a captcha the poster had failed.
- Agents: Perplexity Comet Assistant
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/sanib4r/status/1952765629343166863 (2025-08-05)
- Last verified: 2026-10-16

### Book a restaurant reservation with ChatGPT agent (failure case)
- URL: https://www.realagentcases.com/use-cases/book-a-restaurant-reservation-with-chatgpt-agent-failure-case/
- Outcome: ChatGPT agent took about 5 minutes just to reach the reservation page, then handed back; the poster called it a waste of time.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Took about 5 minutes just to reach the booking page, then handed it back; the poster called it a waste of time.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/AnthonyShoosh/status/1951524727605837921 (2025-08-02)
- Last verified: 2026-10-16

### Weekly meal plan, priced grocery list and spreadsheet with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/weekly-meal-plan-priced-grocery-list-and-spreadsheet-with-chatgpt-agent/
- Outcome: ChatGPT agent created five healthy dinner recipes and a grocery list sorted by store section, checked local prices via Instacart data, and exported everything to Google Sheets.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-tested-chatgpt-agent-on-5-everyday-tasks-heres-what-happened (2025-07-28)
- Last verified: 2026-10-16

### Plan a house-inspection Saturday with ChatGPT agent (blocked)
- URL: https://www.realagentcases.com/use-cases/plan-a-house-inspection-saturday-with-chatgpt-agent-blocked/
- Outcome: ChatGPT agent was asked to search a real-estate site for 4-bedroom Melbourne homes, save listings and route Saturday inspections, but the site and agent blocked each other.
- Agents: ChatGPT agent
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The real-estate site and the agent blocked each other, so the search never happened.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: original
- Source: https://x.com/Hessamsh/status/1949742495539302703 (2025-07-28)
- Last verified: 2026-10-16

### Rename, organize and de-duplicate files with Claude Code
- URL: https://www.realagentcases.com/use-cases/rename-organize-and-de-duplicate-files-with-claude-code/
- Outcome: Claude Code organized and renamed the poster's files and folders and found and removed duplicates; the poster says it worked incredibly well.
- Agents: Claude Code
- Category: Personal
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/JoniYashvili/status/1948767932202238289 (2025-07-25)
- Last verified: 2026-10-16

### Organize 225 Apple Notes with Claude Code
- URL: https://www.realagentcases.com/use-cases/organize-225-apple-notes-with-claude-code/
- Outcome: Claude Code read and changed the local Apple Notes database and organized 225 notes in about 5 minutes with little manual effort.
- Agents: Claude Code
- Category: Personal
- Evidence: Described by poster
- Claimed result: 225 notes organized in about 5 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kirjd/status/1943374929585672439 (2025-07-10)
- Last verified: 2026-10-16

### Build a parenting shift-tracker app with Replit Agent
- URL: https://www.realagentcases.com/use-cases/build-a-parenting-shift-tracker-app-with-replit-agent/
- Outcome: A non-coder built ParentShift, a working app for couples to track night-duty shifts, using Replit Agent; the author reports it took about 3 hours and cost under $50.
- Agents: Replit Agent
- Category: Personal
- Evidence: Screenshots or logs
- Claimed result: Idea to app in 3 hours for $50 (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.producingparadise.com/articles/workflows/from-idea-to-app-in-3-hours-for-50-how-i-used-ai-to-build-parentshift (2025-06-25)
- Last verified: 2026-10-16

## Research

### Report global NAND market share with Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/report-global-nand-market-share-with-gemini-deep-research/
- Outcome: Asked for a simple report on global NAND market share, Gemini Deep Research returned figures the poster called an incredible failure when compared with Sandisk's investor-day slide and Claude's answer to the same query.
- Agents: Gemini Deep Research
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Figures didn't match Sandisk's investor-day data.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/MarceloLima/status/2061791908385419338 (2026-06-02)
- Last verified: 2026-10-16

### Write a one-page sales brief on a target company with Manus
- URL: https://www.realagentcases.com/use-cases/write-a-one-page-sales-brief-on-a-target-company-with-manus/
- Outcome: Asked for a B2B sales brief on Google for a tl;dv pitch, Manus produced a one-page brief with revenue, tech stack, key decision makers, recent announcements and a sourced section mapping Google's pain points to tl;dv features; the reviewer says it took two minutes.
- Agents: Manus
- Category: Research
- Evidence: Article or interview
- Claimed result: About two minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://tldv.io/blog/manus-ai-review/ (2026-04-25)
- Last verified: 2026-10-16

### Extend a family tree two generations with Claude Code and browser agents
- URL: https://www.realagentcases.com/use-cases/extend-a-family-tree-two-generations-with-claude-code-and-browser-agents/
- Outcome: Working from the family's archive scans, Claude Code with memory and the Claude browser plugin cross-checked church books, ship manifests and naturalization records, built a 300-page family wiki, and found two more generations in a day; the family is now checking it for errors.
- Agents: Claude Code, Claude in Chrome, Perplexity Comet
- Category: Research
- Evidence: Shown on video
- Claimed result: 300-page family wiki (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/staskulesh/status/2036173711053651984 (2026-03-23)
- Last verified: 2026-10-16

### Fact-check a viral research-paper claim with Claude in Chrome
- URL: https://www.realagentcases.com/use-cases/fact-check-a-viral-research-paper-claim-with-claude-in-chrome/
- Outcome: Asked to fact-check a post claiming Bitcoin would 'surge 500%' based on a study, Claude for Chrome reported in about 2 minutes that the paper wasn't in any academic database, the 500% maths was wrong, a commenter said it was AI-generated, and the paper itself said it wasn't a prediction.
- Agents: Claude in Chrome
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: About 2 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/seriesfusion/status/2005052460457754963 (2025-12-27)
- Last verified: 2026-10-16

### Find hard-to-locate company tax IDs and addresses with GPT-5 Pro
- URL: https://www.realagentcases.com/use-cases/find-hard-to-locate-company-tax-ids-and-addresses-with-gpt-5-pro/
- Outcome: Working for over 45 minutes, GPT-5 Pro found correct FEINs for 4 of 6 subsidiaries on the Florida Division of Corporations site and one address via a FINRA BrokerCheck report; the poster double-checked the FEINs.
- Agents: GPT-5 Pro
- Category: Research
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/deredleritt3r/status/1986099337303126390 (2025-11-05)
- Last verified: 2026-10-16

### Summarize English and Tamil YouTube videos with Comet
- URL: https://www.realagentcases.com/use-cases/summarize-english-and-tamil-youtube-videos-with-comet/
- Outcome: Comet Assistant summarized English YouTube conversations quickly and well, but on a Tamil discussion it had to play the whole video and made a few mistakes, missing some crucial points.
- Agents: Perplexity Comet Assistant
- Category: Research
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Good on English videos; slow and missed key points on a Tamil one.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/bseshadri/status/1982688143888597156 (2025-10-27)
- Last verified: 2026-10-16

### Test five web tasks with Project Mariner
- URL: https://www.realagentcases.com/use-cases/test-five-web-tasks-with-project-mariner/
- Outcome: In a test of 5 tasks, Project Mariner found YouTube video view counts, but failed to generate relevant video ideas, couldn't preview HTML properly, refused to log into email and couldn't send emails.
- Agents: Project Mariner
- Category: Research
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. 1 of 5 tasks succeeded.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/JulianGoldieSEO/status/1927077282041999503 (2025-05-26)
- Last verified: 2026-10-16

### Find links to content with Project Mariner (no links returned)
- URL: https://www.realagentcases.com/use-cases/find-links-to-content-with-project-mariner-no-links-returned/
- Outcome: Asked to find links to content, Project Mariner found the content but wouldn't return the links, which the poster said made the result not very helpful.
- Agents: Project Mariner
- Category: Research
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Found the content but would not return the links, which the poster said made it unhelpful.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/Indy_triguy/status/1925727943436669316 (2025-05-23)
- Last verified: 2026-10-16

### Research family life in a town you're moving to with Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/research-family-life-in-a-town-you-re-moving-to-with-gemini-deep-research/
- Outcome: Buying a new house, Dan Fitzpatrick ran Gemini Deep Research and ChatGPT Deep Research on family life in the area he was moving to; he says Gemini 'won this by miles' with a wider range of research, more questioning of sources and a more thorough report, delivered as a Google Doc.
- Agents: Gemini Deep Research
- Category: Research
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.linkedin.com/posts/theaieducator_im-in-the-process-of-buying-a-new-house-activity-7306728186598313984-hBxK (2025-03-15)
- Last verified: 2026-10-16

### Complete a Qualtrics survey with OpenAI Operator
- URL: https://www.realagentcases.com/use-cases/complete-a-qualtrics-survey-with-openai-operator/
- Outcome: The poster, a survey researcher, asked Operator to complete a Qualtrics survey they had created; they call the result very promising for Operator but very concerning for survey researchers.
- Agents: OpenAI Operator
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Ingar30/status/1900096859697262976 (2025-03-13)
- Last verified: 2026-10-16

### Get a quick sourced briefing on a local wildlife species with Grok DeepSearch
- URL: https://www.realagentcases.com/use-cases/get-a-quick-sourced-briefing-on-a-local-wildlife-species-with-grok-deepsearch/
- Outcome: Indian Express's reviewer asked Grok 3 DeepSearch about the state of the house sparrow in India; he reports it scoured 101 sources in 46 seconds before presenting its findings.
- Agents: Grok DeepSearch
- Category: Research
- Evidence: Article or interview
- Claimed result: 101 sources in 46 seconds (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://indianexpress.com/article/technology/artificial-intelligence/grok-3-review-elon-musk-chatgpt-ai-model-9848064/ (2025-02-21)
- Last verified: 2026-10-16

### Mine Gmail for on-record PR quotes with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/mine-gmail-for-on-record-pr-quotes-with-claude-cowork/
- Outcome: Connected read-only to Gmail, Claude Cowork found 12 relevant PR pitches in about a minute, then narrowed to eight on-record quotes with publish permission that the journalist verified before use.
- Agents: Claude Cowork
- Category: Research
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://tech.yahoo.com/ai/claude/articles/tried-claude-cowork-gmail-inbox-150900973.html (2026-07-08)
- Last verified: 2026-10-16

### Run async code-research benchmarks with a coding agent
- URL: https://www.realagentcases.com/use-cases/run-async-code-research-benchmarks-with-a-coding-agent/
- Outcome: Asked only about cmarkgfm, Claude Code found six other Python Markdown libraries by itself, benchmarked all seven and wrote a report with charts, in which cmarkgfm came out on top; the author notes he hasn't been reviewing these reports in great detail.
- Agents: Claude Code
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://simonwillison.net/2025/Nov/6/async-code-research/ (2025-11-06)
- Last verified: 2026-10-16

### Compare top social media tools in a spreadsheet with ChatGPT agent and Claude
- URL: https://www.realagentcases.com/use-cases/compare-top-social-media-tools-in-a-spreadsheet-with-chatgpt-agent-and-claude/
- Outcome: Allie K. Miller asked both ChatGPT Agent Mode and Claude's Excel feature to compare top social media tools, and Claude finished in 3 minutes vs 8 with denser decision support, though some prices mixed annual vs monthly.
- Agents: ChatGPT agent, Claude (Excel)
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/alliekmiller/status/1967631471520211354 (2025-09-15)
- Last verified: 2026-10-16

### Make a UK tax research agent from a folder of HMRC manuals with Claude Code
- URL: https://www.realagentcases.com/use-cases/make-a-uk-tax-research-agent-from-a-folder-of-hmrc-manuals-with-claude-code/
- Outcome: Using scraped HMRC manuals and legislation plus a CLAUDE.md that launches parallel subagents, Claude Code answered an ATT exam question in about 5 minutes, earning roughly 2.5 of 3 marks by the author's grading.
- Agents: Claude Code
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://martinalderson.com/posts/building-a-tax-agent-with-claude-code/ (2025-08-21)
- Last verified: 2026-10-16

### Build an Nvidia DCF valuation spreadsheet with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/build-an-nvidia-dcf-valuation-spreadsheet-with-chatgpt-agent/
- Outcome: ChatGPT agent ran for 37 minutes, walked through DCF assumptions and returned a downloadable Excel file with formulas and numbers.
- Agents: ChatGPT agent
- Category: Research
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/alliekmiller/status/1946277369360130465 (2025-07-18)
- Last verified: 2026-10-16

### Build an early-retirement (FIRE) plan with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/build-an-early-retirement-fire-plan-with-chatgpt-agent/
- Outcome: With early access, the poster had ChatGPT agent spend 20 minutes researching Vancouver tax law, modeling savings and FIRE scenarios, and producing a downloadable presentation.
- Agents: ChatGPT agent
- Category: Research
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/rowancheung/status/1945896543263080736 (2025-07-17)
- Last verified: 2026-10-16

### Find and summarize files in Google Drive with Comet Assistant
- URL: https://www.realagentcases.com/use-cases/find-and-summarize-files-in-google-drive-with-comet-assistant/
- Outcome: Comet Assistant kept trying until it found the right files in Google Drive and found a way around opening them, but its summary skipped parts of a PRD after reading only about half.
- Agents: Perplexity Comet Assistant
- Category: Research
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Stopped reading the document about halfway and left out several points from its summary.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/isentropicsoul/status/1944094115819307191 (2025-07-12)
- Last verified: 2026-10-16

### Compile a list of China-tech reporters with Manus
- URL: https://www.realagentcases.com/use-cases/compile-a-list-of-china-tech-reporters-with-manus/
- Outcome: After being pushed to be thorough, Manus produced a downloadable list of 30 journalists covering China tech with their outlets and notable work, though paywalls and captchas got in its way.
- Agents: Manus
- Category: Research
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.technologyreview.com/2025/03/11/1113133/manus-ai-review/ (2025-03-11)
- Last verified: 2026-10-16

### Create an interactive elevator-pitch course with Manus
- URL: https://www.realagentcases.com/use-cases/create-an-interactive-elevator-pitch-course-with-manus/
- Outcome: Manus researched the topic and built a complete multi-module course website on elevator pitching; a second prompt added interactive quizzes and videos.
- Agents: Manus
- Category: Research
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.oneusefulthing.org/p/speaking-things-into-existence (2025-03-11)
- Last verified: 2026-10-16

## Shopping

### Fill an Instacart cart for a risotto recipe with an AI agent: ChatGPT Dots did it, Gemini Spark froze
- URL: https://www.realagentcases.com/use-cases/fill-an-instacart-cart-for-a-risotto-recipe-with-an-ai-agent-chatgpt-dots-did-it/
- Outcome: For PCMag, Ruben Circelli asked three agents to find J. Kenji López-Alt's no-stir risotto recipe, research the best version of each ingredient at the local Wegmans and add them to an Instacart cart. ChatGPT's Dots, once Instacart was connected, found the recipe, researched the ingredients and filled the cart without issue. Meta's Muse got there slowly, needed Circelli to log in and enter a verification code, and added saffron threads that aren't in the recipe. Gemini's Spark found the recipe and proposed ingredients but couldn't access the account, tried to take over Chrome and froze.
- Agents: ChatGPT Dots, Muse, Gemini Spark
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. One agent completed the task; one added an item that wasn't in the recipe and one froze.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://uk.pcmag.com/ai/167960/i-let-3-ai-agents-take-over-my-life-only-1-didnt-fail (2026-10-16)
- Last verified: 2026-10-19

### Find furniture on Facebook Marketplace for a new apartment with Meta's Muse, while Dots got the accounts locked (partial)
- URL: https://www.realagentcases.com/use-cases/find-furniture-on-facebook-marketplace-for-a-new-apartment-with-meta-s-muse-whil/
- Outcome: Scientific American's Mary Randolph was moving apartments and tested Meta's Muse (named Mark) and OpenAI's Dots (named Sam) for a week. Muse logged in to their Facebook account, searched Marketplace for couches, lamps and rugs, and could draft and send messages to sellers with their approval, managing the replies while they did other things. They now own a Muse-sourced dining table. When they gave Dots their Facebook login to do the same, Meta flagged it as a bot and temporarily locked their Instagram, Facebook and WhatsApp. Later, Muse pushed them seven times in three days to join Target's membership program and buy a lamp on sale.
- Agents: Muse, ChatGPT Dots
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Muse found a dining table; logging Dots into Facebook got the poster's Meta accounts temporarily locked.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.scientificamerican.com/article/i-gave-metas-muse-and-openais-dots-my-passwords-for-a-week/ (2026-10-09)
- Last verified: 2026-10-19

### Find the exact replacement part for a broken shaver tab from a photo and a voice note to ChatGPT
- URL: https://www.realagentcases.com/use-cases/find-the-exact-replacement-part-for-a-broken-shaver-tab-from-a-photo-and-a-voice/
- Outcome: A small plastic tab broke on Michal Barodkin's Philips shaver after several years. Rather than replace the shaver, they sent ChatGPT a photo of the broken piece and described the problem in a voice message. It found the replacement part, with its Philips part number and a place to order it. It cost six dollars and they picked it up the next day.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/michal-barodkin_a-small-plastic-tab-broke-on-my-philips-shaver-activity-7513549669239607296-wTRU (2026-10-07)
- Last verified: 2026-10-19

### Send four AI agents to get real car insurance quotes: Gemini and ChatGPT got none, Claude and Muse got five each
- URL: https://www.realagentcases.com/use-cases/send-four-ai-agents-to-get-real-car-insurance-quotes-gemini-and-chatgpt-got-none/
- Outcome: Thy Toeung needed to review their car insurance before their son turned 16. They made a data sheet listing every driver, vehicle and the cover they wanted, and asked Claude, ChatGPT, Gemini and Muse to act as their insurance-shopping agent and get real quotes from six reputable insurers. Gemini's auto-browse explained it couldn't use the websites and handed back a comparison table to fill in. ChatGPT took 2 hours and 279 messages across 12 attempts and got no quotes; Toeung believes sites such as Allstate, Nationwide and Liberty Mutual block AI agents. Claude ran for 3.3 hours and got 5 real quotes. Muse, on its free tier, used several sub-agents and got 5 quotes in about 20 minutes. Toeung switched insurer and got about twice the cover for around 40% of the previous price.
- Agents: Claude, ChatGPT, Gemini, Muse
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/toeung_well-that-was-interesting-i-needed-to-review-activity-7513216577463271438-lPGG (2026-10-06)
- Last verified: 2026-10-19

### Find the cheapest seats left for an opera, a sold-out-looking play and Hamilton with Meta's Muse checking the official sites
- URL: https://www.realagentcases.com/use-cases/find-the-cheapest-seats-left-for-an-opera-a-sold-out-looking-play-and-hamilton-w/
- Outcome: PCWorld's Ben Patterson, a Manhattan theatre lover who can't afford Broadway every night, used Muse to hunt for ticket bargains. Asked how much tickets to Verdi's Macbeth at the Met were, it first gave a general web answer, but when asked to check the official site it opened a browser and found a $71 seat (including a $10 service charge) for a weekday evening, logged in and teed it up for checkout. Patterson cancelled, then had it go through every remaining performance of a limited-run Off-Broadway play, finding which had seats and the best price, $92.41. For Hamilton in the next 30 days it found a rear-mezzanine seat for a Saturday matinee at $222 all-in. Patterson didn't buy this time.
- Agents: Muse
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.pcworld.com/article/3243549/metas-muse-ai-assistant-was-born-to-shop.html (2026-09-25)
- Last verified: 2026-10-19

### Find out which astrophotography gear you actually need before spending a budget, with Claude
- URL: https://www.realagentcases.com/use-cases/find-out-which-astrophotography-gear-you-actually-need-before-spending-a-budget/
- Outcome: tonyarkles was curious about astrophotography and had a budget to fill any gaps in their existing photography kit. Claude's conclusion was to get a $20 intervalometer and a battery adapter that runs off the wall: the lenses and tripod they already had were fine. It also pointed out that Jupiter and Venus were near conjunction, so a dark spot facing west would give a good timelapse that weekend, and later gave rough guidelines for a partial lunar eclipse that produced some amazing photos. tonyarkles now plans to spend the budget on a proper mount, but only after confirming they enjoy the hobby.
- Agents: Claude
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49742552 (2026-09-17)
- Last verified: 2026-10-19

### Buy a child's back-to-school shoes in the right size and a school-approved microphone with Grok Bot
- URL: https://www.realagentcases.com/use-cases/buy-a-child-s-back-to-school-shoes-in-the-right-size-and-a-school-approved-micro/
- Outcome: Asked to buy back-to-school shoes for their kid and a microphone for school, Vekay handed both to Grok Bot. It found a previous size 6 Hoka purchase in their email, opened shoes on the Hoka site and added the chosen pair to the cart. For the microphone (the post later calls it headphones) it checked the school website for requirements and added a couple of options to the Amazon cart; Vekay removed one and checked out. Vekay says what usually takes 30 minutes took 5, while watching TV.
- Agents: Grok Bot
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/rajuvamsi007/status/2091384581387690058 (2026-08-23)
- Last verified: 2026-10-19

### Watch two new phones for a price drop during the price-match window with a daily Claude Cowork check, and claim $40 back
- URL: https://www.realagentcases.com/use-cases/watch-two-new-phones-for-a-price-drop-during-the-price-match-window-with-a-daily/
- Outcome: George Liu bought two Samsung phones from Best Buy, one for themselves and one for their father, and noticed a 30-day price match policy covering drops at Best Buy or major competitors. They spent 15 minutes building a Claude Cowork skill, gave it the model number and told it to check Best Buy, Amazon, Costco and a few other retailers every morning at 7 AM. One morning it found a quiet $20 drop on Best Buy's own product page: not a coupon or a sale, just a lower price. They opened Best Buy's chat, its AI walked them through the price match and a human agent approved a refund of $40 plus tax for the two phones.
- Agents: Claude Cowork
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/george-liu-329a549_ai-claudeai-cowork-activity-7487839192664694784-mf5r (2026-07-28)
- Last verified: 2026-10-19

### Find this week's drugstore deals and coupons to stack with Gemini Spark, though one promo code didn't work (partial)
- URL: https://www.realagentcases.com/use-cases/find-this-week-s-drugstore-deals-and-coupons-to-stack-with-gemini-spark-though-o/
- Outcome: TechCrunch's Sarah Perez asked Gemini Spark for product suggestions for an everyday drugstore trip for household items, based on weekly deals and coupons they could clip. Spark told them which products matching their needs were on sale, suggested coupons to clip in the Walgreens app, and explained how to stack coupons on one item with online promo codes for a pickup order. One of the promo codes was invalid when they tried it, even though they met the requirements Spark had described, but buy-one-get-one-free and rewards deals Spark pointed out made up for it.
- Agents: Gemini Spark
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. Spark found deals and coupons, but one promo code it recommended was invalid.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://techcrunch.com/2026/05/30/i-put-googles-24-7-ai-assistant-gemini-spark-to-work-and-its-actually-pretty-useful/ (2026-05-30)
- Last verified: 2026-10-19

### Work out why new running shoes felt wrong and find the old model second-hand with Gemini
- URL: https://www.realagentcases.com/use-cases/work-out-why-new-running-shoes-felt-wrong-and-find-the-old-model-second-hand-wit/
- Outcome: Kevin R. Kosar replaced a favourite pair of New Balance 880v11 running shoes with a similar-looking $100 Fresh Foam X 860v14, and two test runs left their lower back slightly sore. Kosar asked Gemini to compare the two. It explained that the new shoe is a little heavier and built for overpronators, while the 880v11 is a neutral shoe. Gemini then pointed to second-hand retailers, where Kosar bought two pairs of slightly used 880v11s; the new pair became around-the-house shoes.
- Agents: Gemini
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://beveragesbooksandmore.substack.com/p/how-gemini-helped-me-figure-out-why (2026-04-29)
- Last verified: 2026-10-19

### Hunt for a $400 mixed-metal bracelet on Etsy with ChatGPT Instant Checkout and end with an empty cart (failure)
- URL: https://www.realagentcases.com/use-cases/hunt-for-a-400-mixed-metal-bracelet-on-etsy-with-chatgpt-instant-checkout-and-en/
- Outcome: Business Insider's Jordan Hart tried ChatGPT's new Instant Checkout to find their mom a birthday bracelet in sterling silver and gold, with a $400 budget, starting on Etsy. ChatGPT listed seven Etsy options, but with links rather than a buy button, and one was a $2,400 cuff. Asked to show only items it could sell in the chat, it said it couldn't reliably tell from the product data which were eligible, and Skims and Glossier weren't live yet either. Searching other retailers, it found bracelets from brands such as Kendra Scott and Mejuri that fit the brief, though it had to be asked again for direct links. They didn't love any of them and ended with an empty cart.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Failure. No Instant Checkout-eligible item was found, and the search ended without a purchase.
- Prompt type: Original (as published in the source)
- Source: https://tech.yahoo.com/ai/chatgpt/articles/tried-shopping-gift-using-chatgpt-094202349.html (2025-10-19)
- Last verified: 2026-10-19

### Ask Gemini to order dinner from a delivery app (it refused on an unsupported phone)
- URL: https://www.realagentcases.com/use-cases/ask-gemini-to-order-dinner-from-a-delivery-app-it-refused-on-an-unsupported-phon/
- Outcome: Android Police's Ali Salman Zia asked Gemini on his Pixel 8 to order a chicken burger, then to order it from Foodpanda; both times it gave the same refusal and told him to use a delivery app himself. He found Google limits food and ride ordering to the Pixel 10 series and Galaxy S26, Z Flip 8 and Z Fold 8, for users 18+ in the US or Korea, with supported apps such as DoorDash, Uber Eats and Grubhub.
- Agents: Gemini
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Gemini refused both requests with the same generic message; the feature wasn't available on his phone or for Foodpanda.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/gemini-wont-order-dinner-on-my-phone/ (2026-09-16)
- Last verified: 2026-10-17

### Fill a Walmart pickup cart with 26 meal-plan groceries using ChatGPT computer control, stopping before checkout
- URL: https://www.realagentcases.com/use-cases/fill-a-walmart-pickup-cart-with-26-meal-plan-groceries-using-chatgpt-computer-co/
- Outcome: With school and sports starting, Arthur Sullivan had ChatGPT make a meal plan and grocery list, then logged into Walmart and asked it to arrange the order: pickup time, store brands over name brands, and no purchase without review. Using ChatGPT Work's computer control, it added all 26 items to the cart; the only mistake was one lemon instead of two. Sullivan reviewed the order before buying.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. All 26 items were added, but one quantity was wrong (1 lemon instead of 2); the poster reviewed and placed the order themselves.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/arthurjsullivan_this-weeks-practical-ai-experiment-meal-activity-7497760598911279104-RWvL (2026-08-24)
- Last verified: 2026-10-17

### Hand a family's weekly grocery shop to ChatGPT's in-app browser and find it takes nearly an hour
- URL: https://www.realagentcases.com/use-cases/hand-a-family-s-weekly-grocery-shop-to-chatgpt-s-in-app-browser-and-find-it-take/
- Outcome: Betting they could do the weekly shop in five minutes, Kirk Simmons connected ChatGPT to the family Google Calendar to find a delivery window, gave it a spreadsheet of favourite recipes and household needs, and let its in-app browser shop. It asked what food was left, suggested a meal plan, turned it into a list, chose between two supermarkets on budget and fresh food, and clicked items into the cart one by one. It took nearly an hour, picked a fruit-and-veg box when asked for a fruit box, and spent time comparing ethical chocolate brands; the shop arrived on time and well under budget.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The shop arrived on time and under budget, but took nearly an hour rather than five minutes and included at least one wrong item; the writer wasn't sure they'd do it again.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The writer is a technology lawyer who founded an AI legal-tech company.
- Source: https://www.smh.com.au/business/consumer-affairs/i-bet-my-wife-i-could-do-our-grocery-shop-in-five-minutes-with-ai-i-was-wrong-20260819-p60psl.html (2026-08-23)
- Last verified: 2026-10-17

### Pick back-to-school supplies a daughter actually likes, with Target links, using ChatGPT
- URL: https://www.realagentcases.com/use-cases/pick-back-to-school-supplies-a-daughter-actually-likes-with-target-links-using-c/
- Outcome: A parent gave ChatGPT the school's supply list for her daughter; because it already knew the girl's tastes from past chats, it picked specific items and gave direct Target links. She says her daughter, who usually dislikes whatever she chooses, didn't change a single item, only a few colours, and everything arrived within a few days.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ashleyschendel/status/2089671219775766693 (2026-08-18)
- Last verified: 2026-10-17

### Fill a back-to-school supply cart from the teacher's list with Gemini and Instacart
- URL: https://www.realagentcases.com/use-cases/fill-a-back-to-school-supply-cart-from-the-teacher-s-list-with-gemini-and-instac/
- Outcome: Tom's Guide's Amanda Caswell gave ChatGPT, Gemini and Alexa+ the same school supply list: follow it, stick to required quantities, pick reasonably priced items and skip optional ones. Gemini built an Instacart cart, skipped optional items, asked her to approve colours and totalled about $60-$90 ($105 with the optional water bottle). ChatGPT, on her Walmart account, followed the list carefully but slowly for $92.55; Alexa+ added the optional items and reached $251.48.
- Agents: Gemini, ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/i-let-chatgpt-gemini-and-alexa-shop-my-kids-back-to-school-supplies-only-one-got-it-right (2026-08-08)
- Last verified: 2026-10-17

### Ask ChatGPT, Gemini and Claude for Prime Day tech picks, and check the discounts they claim
- URL: https://www.realagentcases.com/use-cases/ask-chatgpt-gemini-and-claude-for-prime-day-tech-picks-and-check-the-discounts-t/
- Outcome: A TechRadar editor asked ChatGPT, Gemini and Claude what discounted tech to buy on Prime Day. Gemini focused on deals but overstated one: it said Sony WH-1000XM5 headphones were 50% off when they were actually 20% off at $198 (was $248). ChatGPT tailored picks to his interests from past chats but was less tuned in to actual discounts, and Claude, which knew least about him, gave fairly generic deals.
- Agents: ChatGPT, Gemini, Claude
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.techradar.com/ai-platforms-assistants/i-asked-chatgpt-gemini-and-claude-what-discount-tech-to-buy-on-amazon-prime-day-heres-which-ai-gave-the-best-advice (2026-06-23)
- Last verified: 2026-10-17

### Catalogue a DVD collection for resale from shelf photos with ChatGPT
- URL: https://www.realagentcases.com/use-cases/catalogue-a-dvd-collection-for-resale-from-shelf-photos-with-chatgpt/
- Outcome: A How-To Geek writer photographed the spines of hundreds of DVDs and Blu-rays and asked ChatGPT to list them in a spreadsheet with columns useful for resale listings. It built the spreadsheet, correctly classified some PSP discs on the same shelf, and filled in more details on a second pass. He still had to review it by hand: some titles weren't legible, and he warns the AI can hallucinate titles not in the photos.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.howtogeek.com/i-catalogued-hundreds-of-dvds-in-minutes-by-sending-photos-to-chatgpt/ (2026-06-17)
- Last verified: 2026-10-17

### Handle an eBay buyer's 'not as described' partial-refund demand with ChatGPT
- URL: https://www.realagentcases.com/use-cases/handle-an-ebay-buyer-s-not-as-described-partial-refund-demand-with-chatgpt/
- Outcome: On a $500 eBay sale, an overseas buyer asked for a big partial refund, then said the box wasn't factory sealed and sent photos of a scratch. The seller asked ChatGPT what an experienced seller would do; it said don't argue, don't offer a discount, and tell the buyer to open a return through eBay. Because the order used eBay's international shipping program, eBay handled it: the buyer was refunded, the item wasn't returned, and the seller says he kept the money.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/resellerrowe/status/2033217028308521266 (2026-03-15)
- Last verified: 2026-10-17

### Make a Costco weekly grocery list for under $100 with ChatGPT (real prices came to $167)
- URL: https://www.realagentcases.com/use-cases/make-a-costco-weekly-grocery-list-for-under-100-with-chatgpt-real-prices-came-to/
- Outcome: GOBankingRates' Kathryn Jackson asked ChatGPT for a full Costco shopping list for the week under $100; it said the list was for a family of two and totalled $95. On Costco's website the same groceries came to $167, some items only came in larger quantities, and the list left out dairy like milk and cheese, tomatoes, other green vegetables, potatoes and other meats.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT's $95 estimate was $167 at Costco's real online prices, and staples were missing.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.gobankingrates.com/saving-money/food/i-asked-chatgpt-to-make-me-a-costco-weekly-grocery-shopping-list-for-under-100/ (2026-03-08)
- Last verified: 2026-10-17

### Compare contact lens prices across retailers from a prescription with Perplexity and save $150
- URL: https://www.realagentcases.com/use-cases/compare-contact-lens-prices-across-retailers-from-a-prescription-with-perplexity/
- Outcome: Needing new contacts after an eye exam (the old lenses were discontinued), Dmitry Shamis found 1-800-CONTACTS expensive, so gave Perplexity the prescription and asked it to find prices. It checked 1-800-CONTACTS, Costco, Walmart and other sites and organized a comparison; Walmart was $150 cheaper for the same contacts, a retailer Shamis hadn't known sold them.
- Agents: Perplexity
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/dmitryshamis_i-saved-150-on-contacts-yesterday-because-activity-7420101986563518464-g7l7 (2026-01-22)
- Last verified: 2026-10-17

### Find a washing machine deal within budget in a Black Friday sale with Grok
- URL: https://www.realagentcases.com/use-cases/find-a-washing-machine-deal-within-budget-in-a-black-friday-sale-with-grok/
- Outcome: When their old washing machine died, a UAE-based poster asked Grok for help buying a new one within budget. The poster says Grok suggested links, compared options and asked what suited them best, and they got their first-ever Black Friday deal, saving around 600 AED, with delivery due in three days.
- Agents: Grok
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/CPUatOnePercent/status/1995183687772880946 (2025-11-30)
- Last verified: 2026-10-17

### Look for Black Friday laptop deals with ChatGPT (and double-check them)
- URL: https://www.realagentcases.com/use-cases/look-for-black-friday-laptop-deals-with-chatgpt-and-double-check-them/
- Outcome: TechRadar deals writer Alex Whitelock tested ChatGPT on Black Friday laptop deals. Its first picks were pricey and left out MacBooks; asked for budget options, it suggested sub-$200 Chromebooks he felt were too basic. Prices were up to date and brands well known, but after digging he concluded there were better deals elsewhere and says you need to double-check prices and cross-reference expert reviews.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Picks were generic; the tester found better deals elsewhere and had to re-prompt for budget options.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The writer is a deals editor at a site that also publishes deals.
- Source: https://www.techradar.com/ai-platforms-assistants/can-chatgpt-save-shoppers-cash-over-black-friday-heres-my-opinion-as-a-retail-expert (2025-11-18)
- Last verified: 2026-10-17

### Find the right TV in Black Friday sales with ChatGPT's questions and agent
- URL: https://www.realagentcases.com/use-cases/find-the-right-tv-in-black-friday-sales-with-chatgpt-s-questions-and-agent/
- Outcome: Knowing next to nothing about TVs, Tom's Guide's Alex Hughes told ChatGPT (Thinking mode) he wanted a 43- or 50-inch TV for £300 or less and asked it to question him to narrow things down. He then switched to ChatGPT agent to search the sales; it returned options with links, reasons and discount details plus two cheaper alternatives, and a top pick. He says the whole process took 20 minutes, mostly agent search time while he did other things, and left him with 7 suitable deals.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-used-chatgpt-agent-to-find-the-perfect-black-friday-tv-deal-heres-how-you-can-too (2025-11-12)
- Last verified: 2026-10-17

### Find five Prime Day tech gifts under $100 each: ChatGPT vs Claude vs Gemini
- URL: https://www.realagentcases.com/use-cases/find-five-prime-day-tech-gifts-under-100-each-chatgpt-vs-claude-vs-gemini/
- Outcome: TechRadar's editor-at-large gave ChatGPT (GPT-5), Claude (Sonnet 4.5) and Google AI Mode (Gemini 2.5 Pro) the same Prime Day gift brief: five fun tech gifts under $100, smaller than a bread box, ideally heavily discounted. He found all followed the brief closely; Gemini answered fastest and returned the most varied list and the best breakdown, Claude pulled its picks from TechRadar's own guides, and ChatGPT leaned toward Amazon devices.
- Agents: Gemini, ChatGPT, Claude
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.techradar.com/tech/ive-been-shopping-online-for-25-years-but-i-had-no-idea-how-chatgpt-and-gemini-could-change-the-experience (2025-10-08)
- Last verified: 2026-10-17

### Get a full refund for event tickets the seller didn't have, by finding a gap in their terms with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-a-full-refund-for-event-tickets-the-seller-didn-t-have-by-finding-a-gap-in-t/
- Outcome: After buying event tickets from a resale site and finding the seller didn't have them in hand, a buyer was refused a refund. They gave ChatGPT the site's terms of service, and it found they required buyers to expressly accept a delayed-fulfilment model, which they were never asked to do at checkout. With ChatGPT they drafted messages citing this and filed complaints with the BBB, FTC and California Attorney General, and say they had a full refund within days.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/arin-baghermian-39b918123_i-thought-i-was-out-of-luck-on-a-bad-purchase-activity-7363694590203236352-BRgl (2025-08-19)
- Last verified: 2026-10-17

### Order a customised fast-food meal for delivery with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/order-a-customised-fast-food-meal-for-delivery-with-chatgpt-agent/
- Outcome: A Wendy's innovation executive gave ChatGPT agent one prompt to order a customised combo, drink and Frosty from wendys.com for delivery, including his payment info. He says it navigated the site like a real user, hit snags, worked around them and placed the order; a misread CVV field almost broke it.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: The poster works at Wendy's, so has an interest in the result.
- Source: https://www.linkedin.com/posts/mikechorey_genai-freshai-wendystech-activity-7358119584379355137-cusO (2025-08-04)
- Last verified: 2026-10-17

### Try to buy dog treats on Amazon with ChatGPT agent (blocked), then on PetSmart (worked)
- URL: https://www.realagentcases.com/use-cases/try-to-buy-dog-treats-on-amazon-with-chatgpt-agent-blocked-then-on-petsmart-work/
- Outcome: TechRadar's Eric Hal Schwartz asked ChatGPT agent to buy Purina dental chews for his new puppy on Amazon. It hit a 503 error page every time and finally said network restrictions made interacting with Amazon impossible. Asked to buy the same treats at PetSmart, it navigated the store and the multi-step checkout with no problem.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Amazon blocked the agent with repeated 503 errors; the same purchase worked at PetSmart.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.techradar.com/ai-platforms-assistants/chatgpt/i-tried-to-get-chatgpt-agent-and-gemini-to-shop-on-amazon-for-me-but-it-failed-heres-why (2025-08-01)
- Last verified: 2026-10-17

### Buy a work anniversary gift with ChatGPT agent (it abandoned three carts)
- URL: https://www.realagentcases.com/use-cases/buy-a-work-anniversary-gift-with-chatgpt-agent-it-abandoned-three-carts/
- Outcome: A poster asked ChatGPT agent to buy a 4-year work anniversary gift on a $140 budget. It researched and compared well, but abandoned three carts without asking (account required, gift card not found, a form error), unticked marketing emails, made up a phone number that caused shipping issues, chose passwords and rewrote the gift message without consulting him. It finished on a site with guest checkout.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The gift was bought, but the agent made unapproved choices, including an invented phone number that caused shipping problems.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/danielsim_shopify-activity-7356605239640100865-W8-f (2025-07-31)
- Last verified: 2026-10-17

### Make Amazon returns, including one past the deadline, with ChatGPT Operator
- URL: https://www.realagentcases.com/use-cases/make-amazon-returns-including-one-past-the-deadline-with-chatgpt-operator/
- Outcome: A shopper tested ChatGPT's Operator on Amazon returns. The first took about 15 minutes, with the shopper stepping in for login, CAPTCHAs and choices; the second ran autonomously in under a minute but gave a gift-card balance instead of a card refund. For a ~$100 rug past its return window, it needed repeated nudging to reach a human agent instead of messaging the seller and forgot earlier preferences, but in the end got the extension and the $100 back.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/mengmengchen_heres-how-chatgpts-operator-saved-me-100-activity-7292667126014754818-GSFP (2025-02-04)
- Last verified: 2026-10-17

### Build a reusable Gemini shopping-assistant prompt with live product links
- URL: https://www.realagentcases.com/use-cases/build-a-reusable-gemini-shopping-assistant-prompt-with-live-product-links/
- Outcome: A tech columnist asked Gemini 2.0 to write him a shopping-assistant prompt (role, workflow, follow-up questions and web search), loaded it into a chatbot wrapper with web browsing on, and fine-tuned it to return live product links. He says it delivered almost exactly what he needed in early tests, and lets him weight three key features, e.g. 'wet and dry', 'cheap' and 'rotary' for a men's shaver.
- Agents: Gemini
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.tomsguide.com/ai/i-used-gemini-2-0-to-create-an-ai-shopping-assistant-its-surprisingly-good-at-saving-me-time (2025-01-23)
- Last verified: 2026-10-17

### Order a pizza by phone with a coupon using Grok Bot
- URL: https://www.realagentcases.com/use-cases/order-a-pizza-by-phone-with-a-coupon-using-grok-bot/
- Outcome: The poster says Grok Bot found that his pizza shop had moved, set up a call, found a coupon that saved $8.99 and placed the order by phone; he called the shop afterwards to confirm the order was real and shared the call recording.
- Agents: Grok Bot
- Category: Shopping
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/johnbikis/status/2111127651528446333 (2026-10-16)
- Last verified: 2026-10-16

### Clip every digital coupon on a grocery account with Claude Cowork
- URL: https://www.realagentcases.com/use-cases/clip-every-digital-coupon-on-a-grocery-account-with-claude-cowork/
- Outcome: Claude Cowork logged into the poster's Jewel-Osco account and clipped every available coupon.
- Agents: Claude Cowork
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/NYCerInExile/status/2105639156328329221 (2026-10-01)
- Last verified: 2026-10-16

### Rank used cars to test-drive and get a test-drive checklist with Grok
- URL: https://www.realagentcases.com/use-cases/rank-used-cars-to-test-drive-and-get-a-test-drive-checklist-with-grok/
- Outcome: An X user says she gave Grok all the cars she was considering test-driving and it evaluated and ranked them, then gave her a detailed list of what to check on the test drive at each speed and under the hood. She says she bought a one-owner car with 48k miles and no accidents from CarMax.
- Agents: Grok
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/GlamGabber/status/2100408735214805040 (2026-09-17)
- Last verified: 2026-10-16

### Build a grocery cart and check out inside ChatGPT with the Instacart app
- URL: https://www.realagentcases.com/use-cases/build-a-grocery-cart-and-check-out-inside-chatgpt-with-the-instacart-app/
- Outcome: TechRadar's reviewer connected Instacart to ChatGPT and asked for infant formula, vanilla ice cream and fruit; ChatGPT suggested products, built the cart quickly and opened a checkout window in the chat confirming items, payment and address.
- Agents: ChatGPT (Instacart app)
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.techradar.com/ai-platforms-assistants/i-asked-chatgpt-and-claude-to-do-my-grocery-shopping-one-felt-like-the-future (2026-06-23)
- Last verified: 2026-10-16

### Put together an Uber Eats dinner order with Gemini task automation
- URL: https://www.realagentcases.com/use-cases/put-together-an-uber-eats-dinner-order-with-gemini-task-automation/
- Outcome: The Verge's Allison Johnson had Gemini's task automation on a Pixel 10 Pro and Galaxy S26 Ultra build her Saturday dinner order on Uber Eats. It correctly added two half portions to make one chicken teriyaki combo, struggled to find a side of greens and made a couple of wrong turns, and took about nine minutes, stopping before the final confirm so she could check and place the order.
- Agents: Gemini (task automation)
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.theverge.com/tech/898282/gemini-task-automation-uber-doordash-hands-on (2026-03-21)
- Last verified: 2026-10-16

### Put three second-hand jacket options in a Depop cart with Gemini in Chrome
- URL: https://www.realagentcases.com/use-cases/put-three-second-hand-jacket-options-in-a-depop-cart-with-gemini-in-chrome/
- Outcome: Wired's reviewer asked Chrome's Auto Browse for a leather jacket on Depop that fits an XL man; it searched, set the size filter, added three jackets to the cart and wrote a case for each. One was a decent-looking $40 jacket, but he notes it simply picked the top three search results.
- Agents: Gemini in Chrome (auto browse)
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.wired.com/story/google-chrome-auto-browse-hands-on/ (2026-01-30)
- Last verified: 2026-10-16

### Pick a mattress under budget from real reviews with ChatGPT
- URL: https://www.realagentcases.com/use-cases/pick-a-mattress-under-budget-from-real-reviews-with-chatgpt/
- Outcome: The Sydney Morning Herald's Jessica Yun asked ChatGPT for the best mattress around $600, with top three picks based on quality and value informed by real reviews. She looked at all three, chose a hybrid mattress and paid for it in under 10 minutes, $59 under budget, and says she was delighted with it less than two weeks later.
- Agents: ChatGPT
- Category: Shopping
- Evidence: Article or interview
- Claimed result: Bought in under 10 minutes, $59 below budget (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.smh.com.au/business/consumer-affairs/i-used-chatgpt-to-help-me-choose-a-new-mattress-i-didn-t-expect-it-to-go-so-well-20260115-p5nua6.html (2026-01-17)
- Last verified: 2026-10-16

### Order a pizza to your door with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/order-a-pizza-to-your-door-with-chatgpt-agent/
- Outcome: Tom's Guide's Alex Hughes asked ChatGPT agent to order a pizza to his address and ask before paying; it worked through pop-ups and site obstacles and handed back at checkout for his details and payment, taking about 15-20 minutes. Asked for a pizza from one specific restaurant, it gave up on the restaurant's own site, searched Deliveroo at length and finally found it on Uber Eats.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/i-asked-chatgpt-agent-to-order-a-pizza-to-my-house-heres-what-happened (2025-11-01)
- Last verified: 2026-10-16

### Fill a Walmart delivery cart from past purchases with ChatGPT Atlas
- URL: https://www.realagentcases.com/use-cases/fill-a-walmart-delivery-cart-from-past-purchases-with-chatgpt-atlas/
- Outcome: ZDNet's reviewer had Atlas agent mode order wood putty, caulk and screws from her local Walmart for same-day delivery; after she signed in manually and told it to use her past purchases, it found the right brands in her order history and loaded the cart, leaving her to pick a window, tip and pay. About 10 minutes in all.
- Agents: ChatGPT Atlas agent mode
- Category: Shopping
- Evidence: Article or interview
- Claimed result: About 10 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://tech.yahoo.com/ai/chatgpt/articles/let-chatgpt-atlas-walmart-shopping-114600576.html (2025-10-22)
- Last verified: 2026-10-16

### Find an anti-glare screen protector that arrives on time with ChatGPT Atlas
- URL: https://www.realagentcases.com/use-cases/find-an-anti-glare-screen-protector-that-arrives-on-time-with-chatgpt-atlas/
- Outcome: Digital Trends' reviewer asked Atlas agent mode for an anti-glare screen protector for the iPhone 17 Pro; it matched the criteria, chose one that would arrive by his date at his location and stopped at Amazon's checkout page.
- Agents: ChatGPT Atlas agent mode
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.digitaltrends.com/computing/24-hours-in-openais-chatgpt-atlas-ai-browser-is-already-surprising-me/ (2025-10-22)
- Last verified: 2026-10-16

### Plan a meal and order the groceries with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/plan-a-meal-and-order-the-groceries-with-chatgpt-agent/
- Outcome: Sky Canaves tried OpenAI's example 'plan and buy ingredients to make Japanese breakfast for four'; ChatGPT agent shopped her local H-E-B through Instacart, needed her to take over to log in, filled a roughly $130 cart with things she already had, asked before paying and the trimmed order arrived on time. 'It worked, yes', she wrote, but with plenty of friction.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/sky-canaves_i-asked-chatgpt-to-go-grocery-shopping-for-activity-7373436570684620800-hXSu (2025-09-15)
- Last verified: 2026-10-16

### Prep an Instacart order from history and sales with ChatGPT agent (odd picks)
- URL: https://www.realagentcases.com/use-cases/prep-an-instacart-order-from-history-and-sales-with-chatgpt-agent-odd-picks/
- Outcome: Given access to Instacart and told to prep an order from the poster's history and current sales, ChatGPT agent picked items with reasoning the poster found oddly generic ("cereal, which is on sale and a usual part of groceries").
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/osoleve/status/1950032781259477233 (2025-07-29)
- Last verified: 2026-10-16

### Track down a hard-to-find toy and reach checkout with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/track-down-a-hard-to-find-toy-and-reach-checkout-with-chatgpt-agent/
- Outcome: Testing ChatGPT agent for a week, a Tom's Guide writer asked it to find the hard-to-get Big Into Energy Labubu nearby and chose 'purchase' when it asked. In six minutes the agent found it, added it to the cart and went to checkout, then asked for shipping and card details; she took control of the agent's browser to finish the sale.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://tech.yahoo.com/ai/articles/tested-chatgpt-agent-week-good-115833661.html (2025-07-28)
- Last verified: 2026-10-16

### Order groceries from H-E-B with ChatGPT agent (access denied)
- URL: https://www.realagentcases.com/use-cases/order-groceries-from-h-e-b-with-chatgpt-agent-access-denied/
- Outcome: The poster tried to order groceries from H-E-B in Austin with ChatGPT agent, but it failed because access was denied.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Access to the store site was denied, so no order was placed.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/CodyBaker_xx/status/1948746246341943795 (2025-07-25)
- Last verified: 2026-10-16

### Shortlist vintage lamps on Etsy with ChatGPT agent (cart claim was false)
- URL: https://www.realagentcases.com/use-cases/shortlist-vintage-lamps-on-etsy-with-chatgpt-agent-cart-claim-was-false/
- Outcome: On ChatGPT agent's first day, The Verge's Hayden Field ran its suggested task to find five Japanese-inspired vintage lamps on Etsy under $200 with free shipping and add them to her cart. It worked for 50 minutes and then said it had added five lamps totalling around $825 to her cart, but when she checked Etsy on her own computer the cart was empty, because the agent was using its own browser session.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. After 50 minutes it claimed five lamps were in her cart, but her Etsy cart was empty.
- Prompt type: Original (as published in the source)
- Context: Prompt: ChatGPT's own suggested example task, which the writer clicked to fill in, not the writer's own wording.
- Source: https://www.theverge.com/ai-artificial-intelligence/710020/openai-review-test-new-release-chatgpt-agent-operator-deep-research-pro-200-subscription (2025-07-18)
- Last verified: 2026-10-16

### Turn 200 photos of household items into a moving-sale list with Claude Code
- URL: https://www.realagentcases.com/use-cases/turn-200-photos-of-household-items-into-a-moving-sale-list-with-claude-code/
- Outcome: Claude Code processed over 200 photos of the poster's household items and created a moving-sale list with descriptions and prices; the poster says it saved hours compared with writing Facebook Marketplace listings by hand.
- Agents: Claude Code
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/rautility/status/1937172819215065456 (2025-06-23)
- Last verified: 2026-10-16

### Compare prices for 60–70 garden plants across 5 sites with OpenAI Operator
- URL: https://www.realagentcases.com/use-cases/compare-prices-for-60-70-garden-plants-across-5-sites-with-openai-operator/
- Outcome: Overnight, Operator cross-referenced a list of 60–70 plants across 5 sites into a price comparison table; the poster called it buggy but fun.
- Agents: OpenAI Operator
- Category: Shopping
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/b1rdmania/status/1931618971968544834 (2025-06-08)
- Last verified: 2026-10-16

### Find a store with a chair in stock using OpenAI Operator
- URL: https://www.realagentcases.com/use-cases/find-a-store-with-a-chair-in-stock-using-openai-operator/
- Outcome: The only shop the poster had found had 60 working days' delivery for a chair, so they asked Operator to find a store with it in stock; 11 minutes later it had.
- Agents: OpenAI Operator
- Category: Shopping
- Evidence: Described by poster
- Claimed result: 11 minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Jobvo/status/1930684200203567287 (2025-06-05)
- Last verified: 2026-10-16

### Research the best car to buy for $20k with Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/research-the-best-car-to-buy-for-20k-with-gemini-deep-research/
- Outcome: The Neuron asked Gemini Deep Research for the best car to buy for $20k; it says Gemini researched 481 websites and added a lifetime ownership cost comparison of a Tesla Model 3, Toyota Prius and Camry.
- Agents: Gemini Deep Research
- Category: Shopping
- Evidence: Article or interview
- Claimed result: 481 websites researched (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.theneurondaily.com/p/gpt-deep-research-tested (2025-02-04)
- Last verified: 2026-10-16

### Find working coupon codes at checkout with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-working-coupon-codes-at-checkout-with-chatgpt-agent/
- Outcome: ChatGPT agent kept trying coupon codes until it found ones that worked; the poster says it saved them $79.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Screenshots or logs
- Claimed result: $79 saved (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/joevkolo/status/1976496831623463109 (2025-10-10)
- Last verified: 2026-10-16

### Find local takeout deals under a budget with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-local-takeout-deals-under-a-budget-with-chatgpt-agent/
- Outcome: The Tom's Guide author uses ChatGPT agent to scour delivery apps and local restaurants for family takeout deals under a set budget.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.tomsguide.com/ai/5-clever-ways-i-use-chatgpt-agent-to-save-serious-money-each-month (2025-10-10)
- Last verified: 2026-10-16

### Daily product restock check with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/daily-product-restock-check-with-chatgpt-agent/
- Outcome: A scheduled ChatGPT agent checks every day whether a product is in stock, reports at 8am, and even flags status changes like 'out of stock' becoming 'coming soon'.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/starsailing11/status/1969125311565430944 (2025-09-19)
- Last verified: 2026-10-16

### Order weekly meal delivery with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/order-weekly-meal-delivery-with-chatgpt-agent/
- Outcome: ChatGPT agent logged into Gym Meals Direct, chose meals and sizes, picked Monday delivery and applied a discount.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/benny_alexander/status/1952206488023585198 (2025-08-04)
- Last verified: 2026-10-16

### Draft and complete an eBay Germany listing from photos with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/draft-and-complete-an-ebay-germany-listing-from-photos-with-chatgpt-agent/
- Outcome: From a few photos, ChatGPT agent identified an Apple Watch Series 7, noted display scratches, drafted a German title and description, then finished the eBay DE listing draft after login in about 15 minutes.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/rschu/status/1950470327252705510 (2025-07-30)
- Last verified: 2026-10-16

### Request a furniture quote with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/request-a-furniture-quote-with-chatgpt-agent/
- Outcome: In 11 minutes ChatGPT agent ran 13 searches, skipped JavaScript-heavy sites it couldn't read, and shortlisted wicker daybeds from server-rendered retailers so it could request a quote in Sydney.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.leannewong.co/chatgpt-agent-ai-agentic-seo/ (2025-07-24)
- Last verified: 2026-10-16

### Fill an online grocery cart from a handwritten list with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/fill-an-online-grocery-cart-from-a-handwritten-list-with-chatgpt-agent/
- Outcome: ChatGPT agent transcribed 15 of 16 items from a photographed shopping list (missing onions), found the nearest Harris Teeter and added them to the cart; a safety monitor first blocked the Kroger login, but after the user said they'd log in themselves and entered the password, it ordered the groceries.
- Agents: ChatGPT agent
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Missed one of 16 items and gave up at the store login until the user stepped in; the author called it still not very useful.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://www.understandingai.org/p/chatgpt-agent-a-big-improvement-but (2025-07-22)
- Last verified: 2026-10-16

### Order groceries from a 16-item list with OpenAI Operator
- URL: https://www.realagentcases.com/use-cases/order-groceries-from-a-16-item-list-with-openai-operator/
- Outcome: Operator misread parts of a 16-item grocery list and added only 13 items to the cart even after corrections; the author concluded it wasn't ready for real use.
- Agents: OpenAI Operator
- Category: Shopping
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.understandingai.org/p/computer-use-agents-seem-like-a-dead (2025-06-26)
- Last verified: 2026-10-16

## Travel

### Ask airlines to refund late-night rides after delayed flights with Meta's Muse, which drafted good emails but missed that airlines don't take email (partial)
- URL: https://www.realagentcases.com/use-cases/ask-airlines-to-refund-late-night-rides-after-delayed-flights-with-meta-s-muse-w/
- Outcome: Gizmodo's Webb Wright had some badly delayed flights that forced them to pay for Lyft rides home in the early hours, and asked Muse to request refunds. Given screenshots of the receipts and the flight details, Muse drafted decent emails, courteous but firm. But buried inside one draft was the important point that the airlines didn't take refund requests by email, so the drafts would have to be pasted into the airlines' websites. Webb ended up opening a browser and entering the details by hand, though they did use the message Muse had drafted.
- Agents: Muse
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Good drafts, but the submission method was buried and the poster had to file by hand.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://gizmodo.com/i-asked-muse-to-help-me-book-my-holiday-travel-it-didnt-fly-far-2000823264 (2026-10-12)
- Last verified: 2026-10-19

### Book a trip, flight, hotel and shuttle with ChatGPT Dots, and keep confirming the same booking (partial)
- URL: https://www.realagentcases.com/use-cases/book-a-trip-flight-hotel-and-shuttle-with-chatgpt-dots-and-keep-confirming-the-s/
- Outcome: neom spent two hours trying ChatGPT's Dots agent and grades it B-. Asked to figure out a trip, it did; neom booked the flight and hotel, and Dots noticed the shuttle hadn't been booked and offered to book it with the details they'd discussed. But after neom said yes, it went back to the site, re-checked the details, asked for confirmation, filled in the booking form and then asked again whether to book. It can sign in with passwords, but wouldn't fetch two-factor codes from email even with email access, and its permissions checker rejected broad rules neom wrote, such as proceeding with routine, reversible actions without asking. neom finds it faster to use regular chat.
- Agents: ChatGPT Dots
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The trip got booked, but the agent kept re-asking for confirmation and couldn't handle sign-in codes.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49904198 (2026-09-30)
- Last verified: 2026-10-19

### Find two seven-seat rental cars far cheaper through a reseller discount with Codex searching every membership and site
- URL: https://www.realagentcases.com/use-cases/find-two-seven-seat-rental-cars-far-cheaper-through-a-reseller-discount-with-cod/
- Outcome: Xcelerate needed two rental cars that could each carry seven people for a trip to Maine, and had memberships with Avis, Hertz, Costco and Chase's travel portal; checking every combination by hand would have taken forever. They wrote a Codex prompt in 20 minutes and let it use the browser directly. It queued any CAPTCHAs for them to solve, and a sub-agent searched Reddit for other people's ways of getting good rental deals. It found a third-party Avis reseller charging $600 with a discount code, against $1,100 cheapest from Avis directly. By those quoted prices that's about $500 less for both cars (the poster calls it a $1,000 saving), and they say they wouldn't have found it without about a week of manual searching.
- Agents: Codex
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=49899460 (2026-09-29)
- Last verified: 2026-10-19

### Get about 70% of a visa application done with Meta's Muse agent, then finish the uploads and one flight by hand (partial)
- URL: https://www.realagentcases.com/use-cases/get-about-70-of-a-visa-application-done-with-meta-s-muse-agent-then-finish-the-u/
- Outcome: Ha Tran Nguyen Phuong used Muse to apply for a visa. It booked a refundable multi-leg flight and a refundable hotel, got the bank statement, dug through email for papers such as the passport photo and I-94, wrote a day-by-day itinerary and filled in the online form. But it couldn't size the passport photo (made from a selfie with ChatGPT) to fit the frame or upload most of the files; one field accepted only JPG, not PDF, and images had to be under 2MB, so they had Muse convert and compress files. Muse also missed extra documents and flights the instructions required, and bot detection stopped it booking one flight. Muse did about 70% of the work and they still spent about an hour on it.
- Agents: Muse
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Muse did about 70% of the work; uploads, photo sizing and one booking had to be done by hand.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/sherlockieee/status/2099371464969589123 (2026-09-14)
- Last verified: 2026-10-19

### Get real flight and hotel prices from a Claude Code travel skill, which only guessed price ranges until it was given a browser (partial)
- URL: https://www.realagentcases.com/use-cases/get-real-flight-and-hotel-prices-from-a-claude-code-travel-skill-which-only-gues/
- Outcome: Christopher Johnson built a Claude Code skill in about twenty minutes to act as a travel agent, sorting flights and places to stay for a business trip. Asked for real prices, it gave ranges, such as 'Flights typically run $280 to $420', and none of them was a price they could actually pay. Claude explained that the booking sites load prices with JavaScript after the page renders, which its tools couldn't see, so it had been estimating. With the Claude in Chrome extension installed and the same skill and request, it opened the sites and returned three specific places to stay with current prices and direct links. It was slower and used more of the plan's usage limits. They kept 'Act without asking' off, and Claude never bought anything.
- Agents: Claude Code, Claude in Chrome
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Without a browser the skill gave guessed ranges; with Claude in Chrome it gave real options.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://cjonsystems.substack.com/p/my-ai-travel-agent-was-guessing (2026-08-26)
- Last verified: 2026-10-19

### Check a visa rule with a quick Google Gemini answer, and get denied boarding to Honduras
- URL: https://www.realagentcases.com/use-cases/check-a-visa-rule-with-a-quick-google-gemini-answer-and-get-denied-boarding-to-h/
- Outcome: Daniel Ross, a UK citizen travelling through Central America, was in Medellín when they realised they hadn't checked whether Honduras needed a visa. They searched in Chrome on their phone, and Google's Gemini answer said US, Canadian, UK and EU citizens don't need a visa for tourist stays of up to 90 days. They took its word for it. At the airport the agent asked for a visa, and the UK government's travel page showed one was needed, obtainable only at a Honduran embassy. Airline staff helped them rebook to end in Panama instead. The mistake cost $457, and Ross says they will stick to official sources for visa rules.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Failure. The AI answer said no visa was needed; one was, and the traveller was denied boarding.
- Prompt type: Original (as published in the source)
- Source: https://upgradedpoints.com/news/ai-visa-advice-denied-boarding/ (2026-08-06)
- Last verified: 2026-10-19

### Rebuild a delayed-bag claim with Claude: the full email history for a new agent, and a receipts spreadsheet
- URL: https://www.realagentcases.com/use-cases/rebuild-a-delayed-bag-claim-with-claude-the-full-email-history-for-a-new-agent-a/
- Outcome: Manuel Dominguez's suitcase never left Canada on a WestJet trip to Mexico City and Puerto Vallarta, and it took 10 days to get it back. When the case moved from one WestJet person to another, they asked Claude to pull the full history of emails with the first person so they could pass it on. To claim expenses, they photographed the receipts for clothes bought during those 10 days and asked Claude to build an Excel sheet from them to attach to the claim. It saved a lot of time, but WestJet paid only 60% of the original claim.
- Agents: Claude
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. Claude saved time on the paperwork, but WestJet paid only 60% of the claim.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/manuelmdz_i-finally-got-payment-for-a-delayed-bag-from-activity-7485111103044325376-9sUG (2026-07-20)
- Last verified: 2026-10-19

### Price the same Tokyo hotel stay through four booking sites to the final checkout page with Claude Code agents, and book the cheapest
- URL: https://www.realagentcases.com/use-cases/price-the-same-tokyo-hotel-stay-through-four-booking-sites-to-the-final-checkout/
- Outcome: Kin Meng Sio launched 4 Claude Code agents to go through 4 booking flows for the same Tokyo hotel and dates (no breakfast, June 13–15) all the way to the final confirmation page, with coupons applied, taxes added and currency holds calculated. It took five minutes while the agents controlled their browsers; it used to take an hour of tab-switching. Booking direct would have cost $93 more than Agoda and had the worst cancellation policy of the four: sliding 40/50/100% fees within 3 days, while every online travel agency offered free cancellation. They booked Agoda at $192.27. Earlier the same week, the AI had caught a $702 drop fee on a car rental.
- Agents: Claude Code
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/kinsio_claude-just-booked-my-hotel-stay-i-saved-activity-7461082996692000768-c92y (2026-05-15)
- Last verified: 2026-10-19

### Plan a four-day family trip to London with ChatGPT, then have hotel concierges catch what it got wrong (partial)
- URL: https://www.realagentcases.com/use-cases/plan-a-four-day-family-trip-to-london-with-chatgpt-then-have-hotel-concierges-ca/
- Outcome: Carley Thornell, a working parent, spent a few hours going back and forth with ChatGPT to plan a four-day London trip with their 6-year-old, ending up with a spreadsheet itinerary. On the ground, hotel concierges caught what it missed: a V&A exhibit sold out since day one with no walk-ups, a way into the British Museum's Egypt wing when general tickets were gone, a nut-free restaurant pick (Ping Pong) that had closed the previous year despite a 'permanently closed' flag on Google, and suggested bus routes 11 and 15 that don't go to the Science Museum or Hamleys. A rail strike the week they were there wasn't something ChatGPT could know. Some of its dining picks still worked out for the family.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Partial. The plan worked as a draft, but concierges caught a closed restaurant, wrong bus routes and sold-out exhibits.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://10best.usatoday.com/travel/chatgpt-planned-my-london-trip/ (2026-03-18)
- Last verified: 2026-10-19

### Plan a sunset hike from a ChatGPT itinerary and get stuck at the summit after the last ropeway had gone (failure)
- URL: https://www.realagentcases.com/use-cases/plan-a-sunset-hike-from-a-chatgpt-itinerary-and-get-stuck-at-the-summit-after-th/
- Outcome: Dana Yao and their spouse used ChatGPT to plan a romantic hike up Mount Misen on Japan's Itsukushima island. They set off at 15:00 to reach the summit for sunset, exactly as ChatGPT had told them. ChatGPT said the last ropeway down was at 17:30, but when they got to the station to descend, the ropeway had already closed, and they were stuck at the top of the mountain.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Failure. ChatGPT gave the wrong last ropeway time and the couple were stranded at the summit.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.bbc.com/travel/article/20250926-the-perils-of-letting-ai-plan-your-next-trip (2025-09-29)
- Last verified: 2026-10-19

### Underpack for a Mount Shasta climb after following Gemini's advice, ending in a rescue
- URL: https://www.realagentcases.com/use-cases/underpack-for-a-mount-shasta-climb-after-following-gemini-s-advice-ending-in-a-r/
- Outcome: According to the U.S. Forest Service, three novice hikers planned a Mount Shasta climb partly with Google Gemini, which recommended supplies for an 8-hour climb rather than the 16 hours it took and told them to eat simple carbohydrates rather than fats. They reached the summit at 7 p.m., got lost off-route on the way down after their phone battery died, and one hiker injured a knee at about 11,500 feet. Rangers reached them the next morning; the hikers said they 'relied too much on AI rather than our own critical thinking.'
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Officials said the AI advice underestimated the climb by half and the group needed a ranger rescue after an injury.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Second-hand account: the details come from Forest Service statements reported by the station, not from the hikers directly.
- Source: https://www.kcra.com/article/roseville-hikers-rescued-mount-shasta-ai-advice/73607326 (2026-09-03)
- Last verified: 2026-10-18

### Plan a six-day Oregon trip day by day with ChatGPT and Google Maps
- URL: https://www.realagentcases.com/use-cases/plan-a-six-day-oregon-trip-day-by-day-with-chatgpt-and-google-maps/
- Outcome: The poster booked flights and Airbnbs, then for each day gave ChatGPT the start and end Airbnb, the kind of places they like and how much walking they were up for, and asked what to see along the route. They checked every suggestion on Google Maps, asked for hidden spots, then asked whether each day was feasible and what they'd missed, which turned up a redwood trail. Planning took about four hours; the poster says the plan worked almost perfectly without feeling rushed.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/teja2495/status/2089046423425716386 (2026-08-16)
- Last verified: 2026-10-18

### Win a Lufthansa refund for a cancelled trip over months of ChatGPT-written emails
- URL: https://www.realagentcases.com/use-cases/win-a-lufthansa-refund-for-a-cancelled-trip-over-months-of-chatgpt-written-email/
- Outcome: After a Lisbon–Frankfurt–Miami flight was delayed, boarded and then cancelled, the poster gave up the trip and flew elsewhere. Lufthansa tried to charge for the Lisbon–Frankfurt leg and offered only 'the difference'. The poster fought it entirely by email using ChatGPT, and after months and a long chain of emails got the refund. They still lost money, days and the trip with their mother.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. The refund came through after months, but the poster still lost money, days and the trip itself.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/oracles/status/2058948657253429538 (2026-05-25)
- Last verified: 2026-10-18

### Draft a 13-day cross-country motorcycle route with Grok and ChatGPT, then catch a badly wrong route map
- URL: https://www.realagentcases.com/use-cases/draft-a-13-day-cross-country-motorcycle-route-with-grok-and-chatgpt-then-catch-a/
- Outcome: Planning a 13-day cross-country motorcycle ride for a year ahead, the poster asked Grok and ChatGPT for an initial route and itinerary. Grok's itinerary was 'a little wonky'; ChatGPT's schedule and route looked great. But when asked to draw a map of the route it had described, ChatGPT produced one that the poster says gets worse the more you look at it, making them question whether the itinerary itself was accurate.
- Agents: ChatGPT, Grok
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT's written itinerary looked good to the poster, but the map it generated of the same route was badly wrong; Grok's itinerary was 'a little wonky'.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/RVA4n6/status/2031003709442830644 (2026-03-09)
- Last verified: 2026-10-18

### Avoid the wrong travel insurance for a student abroad with ChatGPT
- URL: https://www.realagentcases.com/use-cases/avoid-the-wrong-travel-insurance-for-a-student-abroad-with-chatgpt/
- Outcome: The poster's son, a Canadian studying in the U.S., was travelling overseas, and the poster almost bought travel insurance out of habit. ChatGPT pointed out the key issue: for this cover, residency mattered more than citizenship. They compared options and chose IMG Patriot International.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/JaneOnAI/status/2027530285328531619 (2026-02-27)
- Last verified: 2026-10-18

### Find the best Starbucks detour on a family drive with Grok in a Tesla
- URL: https://www.realagentcases.com/use-cases/find-the-best-starbucks-detour-on-a-family-drive-with-grok-in-a-tesla/
- Outcome: About 90 minutes from their destination, with the family wanting hot chocolate and coffee, Shiva Arunachalam asked the Grok integration in their Model Y to find the best Starbucks on the way, with the shortest detour, the right direction of travel and timing so drinks would be ready, and to adjust the route. Arunachalam says it picked the right store and re-routed cleanly; they placed the mobile order by hand in the Starbucks app, and the drinks were ready on arrival.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Comments on the post point out that ordinary map apps can also add stops along a route.
- Source: https://www.linkedin.com/posts/arunash_appliedai-autonomy-activity-7413691386673393664-mBRq (2026-01-04)
- Last verified: 2026-10-18

### Plan a romantic, off-the-beaten-path Italy trip when you're too burned out to plan with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-romantic-off-the-beaten-path-italy-trip-when-you-re-too-burned-out-to-pla/
- Outcome: Burned out by work and a London heatwave, Santiago Barraza Lopez and their partner handed the planning to ChatGPT with one prompt and three rules: affordable, romantic, away from the tourist crowds. Its picks, all tied to a cheap flight into Salerno, were Salerno's old town, the fishing village of Cetara and Atrani in the shadow of Amalfi. Barraza Lopez says letting go of the planning brought calm and connection, and they came home feeling recalibrated.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.businessinsider.com/burned-out-asked-chatgpt-plan-trip-italy-2025-10 (2025-10-05)
- Last verified: 2026-10-18

### Find off-the-radar stops along a 5,600 km road trip with ChatGPT o3
- URL: https://www.realagentcases.com/use-cases/find-off-the-radar-stops-along-a-5-600-km-road-trip-with-chatgpt-o3/
- Outcome: Planning a week-long road trip from Chicago to San Francisco via Alabama, Mississippi and Louisiana, the poster gave ChatGPT o3 a rough route, the things they wanted to see and questions in XML format, then refined the daily plans and switched part of the route. Asking 'is there anything to see between [x] and [y]' surfaced places they'd never heard of, including the Space and Rocket Center in Huntsville, Dauphin Island and its car ferry, and Saguaro and Guadalupe Mountains national parks, which they count among the best parts of the trip.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster works at an AI company (Anthropic) and was describing a different company's chatbot.
- Source: https://x.com/gitpushoriginm8/status/1955009180311707972 (2025-08-11)
- Last verified: 2026-10-18

### Get the lay of the land for a trip in half an hour and book a rental the same night with ChatGPT (partial)
- URL: https://www.realagentcases.com/use-cases/get-the-lay-of-the-land-for-a-trip-in-half-an-hour-and-book-a-rental-the-same-ni/
- Outcome: Business Insider's Kelsey Vlamis usually spends hours reading about a destination before booking. For a Door County, Wisconsin trip, Vlamis asked ChatGPT which attractions people call overrated and which are worth the crowds, for timing tips, varied sample itineraries and which town suits someone who likes good food, nature, vintage shops and fewer crowds. In about half an hour Vlamis felt confident enough to book a waterside rental that same night. ChatGPT also recommended at least two restaurants that had closed years earlier.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. ChatGPT recommended at least two restaurants that had closed several years earlier, and the writer still planned to do some reading for hidden gems.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.businessinsider.com/travel-chatgpt-use-case-plan-trips-saves-hours-2025-7 (2025-07-19)
- Last verified: 2026-10-18

### Plan a parent-and-daughter road trip from the Bay Area to Texas, with a halfway stop and hotels, using Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-parent-and-daughter-road-trip-from-the-bay-area-to-texas-with-a-halfway-s/
- Outcome: A parent asked Grok on an iPhone to plan a road trip with their daughter from the Bay Area to Tesla Giga Texas, with a stop halfway and hotel recommendations; then booked the hotels and added them to the calendar. On the road they asked Grok for food and rest stops when needed. The poster says LA traffic added 7 hours to one day, and they missed the factory tour because tickets sold out.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/AlSultan_Meriam/status/2110205309256143333 (2026-10-14)
- Last verified: 2026-10-17

### Get a $1,373.97 refund for a changed flight via a bank chargeback, on Grok Bot's advice
- URL: https://www.realagentcases.com/use-cases/get-a-1-373-97-refund-for-a-changed-flight-via-a-bank-chargeback-on-grok-bot-s-a/
- Outcome: After an airline schedule change, a traveller asked for a full refund on two China Eastern tickets booked through an online travel agency; the agency confirmed it in writing on Sep 8 but then stalled through 8 complaint emails. Grok Bot advised him to stop waiting and dispute the charge with his bank. On Oct 11 the bank ruled in his favour and he got the full $1,373.97 back.
- Agents: Grok Bot
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/weird_ceo/status/2109664238818164876 (2026-10-12)
- Last verified: 2026-10-17

### Plan a Hong Kong trip with Gemini, Grok and ChatGPT (some addresses were made up)
- URL: https://www.realagentcases.com/use-cases/plan-a-hong-kong-trip-with-gemini-grok-and-chatgpt-some-addresses-were-made-up/
- Outcome: A traveller planned a Hong Kong trip mainly with Gemini, using Grok and ChatGPT as extras, to organise locations, build itineraries, make packing lists and turn store addresses into Google Maps links. The traveller says routes were logical and personalised, but the AIs sometimes produced convincing addresses that didn't exist, and got names, opening hours or whether shops were still open wrong. Next time they'd verify every important detail on maps and official sites.
- Agents: Gemini, Grok, ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Itineraries were useful, but some addresses, hours and shop details were made up or wrong.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/raychenAI/status/2106382000877908214 (2026-10-03)
- Last verified: 2026-10-17

### Plan a flexible beach day with an infant in Goa by voice with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-flexible-beach-day-with-an-infant-in-goa-by-voice-with-gemini/
- Outcome: Overwhelmed by pins on Google Maps while travelling with an infant, Android Police's Akshay Bhalla asked Gemini by voice for a flexible day at Mandrem Beach in North Goa: sunset, local food at a beach shack rated 4.2 or higher, maybe water sports, using maps, weather, UV index, traffic and blogs. Gemini suggested a creek in the morning, a local restaurant for lunch, shade during the peak UV hours of 11:30 AM to 3:00 PM, and the beach from 5:30 PM for sunset, then added a café with good Wi-Fi when asked.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/gemini-google-maps-vacation-comparison/ (2026-09-26)
- Last verified: 2026-10-17

### Get back to the US without a passport book after being denied boarding, with Grok's plan
- URL: https://www.realagentcases.com/use-cases/get-back-to-the-us-without-a-passport-book-after-being-denied-boarding-with-grok/
- Outcome: Denied a flight back to the US because they only had a passport card, not the book, a traveller asked Grok what to do. The traveller says it suggested taking an Uber to the border, linked them with a private car from the border to the airport, said which security checkpoint to use and calculated the travel time; they missed the first flight but caught the layover after a 4-hour drive.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/HerroCrypto/status/2100771365674746362 (2026-09-18)
- Last verified: 2026-10-17

### Get a refund from Airbnb for a stay with no hot water using ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/get-a-refund-from-airbnb-for-a-stay-with-no-hot-water-using-chatgpt-agent/
- Outcome: After an Airbnb stay with no hot water and a host-required third-party app that didn't work, a guest explained the problem to ChatGPT by voice and uploaded a video. Instead of drafting an email, ChatGPT opened its cloud browser, had the guest log in to Airbnb, found the stay, opened a support ticket, asked for a human, made the case and uploaded screenshots from the video; the guest only stepped in to approve messages. A representative called and the guest got $250 back.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/robinlobo_chatgpt-got-me-a-250-refund-from-airbnb-activity-7504553893784076288-4szN (2026-09-12)
- Last verified: 2026-10-17

### Plan a month of family travel and remote work across Europe with Gemini, from route to co-working bookings
- URL: https://www.realagentcases.com/use-cases/plan-a-month-of-family-travel-and-remote-work-across-europe-with-gemini-from-rou/
- Outcome: Planning a month of travel with kids while working remotely, Nick Thomson gave Gemini the family's profile, preferences and goals. It talked them out of a there-and-back route from London to southern Italy, suggesting three routes starting with a flight to the furthest point; recommended beaches by the kids' ages and swimming ability, off-peak trains, surf schools and teen-friendly activities; flagged when train companies release cheap tickets; and found co-working spaces, compared day rates and drafted booking emails in French and Spanish. Over the month, one recommendation was wrong: a bus to a market that didn't exist.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ngthomson_the-first-question-everyone-asks-about-our-activity-7497964037176168448-_qkz (2026-08-25)
- Last verified: 2026-10-17

### Use Gemini each morning on holiday to plan the day's route, sights and meals with Google Maps
- URL: https://www.realagentcases.com/use-cases/use-gemini-each-morning-on-holiday-to-plan-the-day-s-route-sights-and-meals-with/
- Outcome: On vacation, Stefan Heil asked Gemini every morning for a route for the day: what to see, where to eat and roughly when to be where. Through its Google Maps integration it surfaced places they wouldn't have found, sanity-checked their own plans, and found a restaurant and somewhere to unwind at the end of long days. Heil notes they didn't compare it with other AI tools.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/stefanheil_googlegemini-ai-traveltech-activity-7493905578424025090-rKI9 (2026-08-14)
- Last verified: 2026-10-17

### Build a calm, step-by-step Osaka trip guide for a neurodivergent traveller with ChatGPT
- URL: https://www.realagentcases.com/use-cases/build-a-calm-step-by-step-osaka-trip-guide-for-a-neurodivergent-traveller-with-c/
- Outcome: A traveller living with depression says ChatGPT helped them build a personal Osaka travel project: two books designed for them as a neurodivergent traveller, the second a field guide they planned to test in Osaka that September. They say it made the trip feel possible.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written before the trip; the result isn't reported.
- Source: https://x.com/hopeforwh/status/2079466954885370272 (2026-07-21)
- Last verified: 2026-10-17

### Read signs, menus and supermarket labels abroad with Gemini Live
- URL: https://www.realagentcases.com/use-cases/read-signs-menus-and-supermarket-labels-abroad-with-gemini-live/
- Outcome: On a trip to Barcelona, Android Police's Rajesh Pandey pointed Gemini Live's camera at Spanish signs and asked what they meant, had it find vegetarian lasagna and other vegetarian ready meals on a supermarket shelf, and used it to understand local café menus instead of ordering the one dish he recognized. He still used Google Translate's Live Conversation for talking with locals and hospital staff.
- Agents: Gemini Live
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/gemini-live-the-travel-assistant-i-didnt-know-i-needed/ (2026-06-21)
- Last verified: 2026-10-17

### Plan a two-week Taiwan and Hong Kong trip with a toddler using Gemini and Ask Maps
- URL: https://www.realagentcases.com/use-cases/plan-a-two-week-taiwan-and-hong-kong-trip-with-a-toddler-using-gemini-and-ask-ma/
- Outcome: A New York Times personal tech columnist used Gemini (with personal data from Gmail) and Google Maps' Ask Maps to plan a 14-day Taiwan and Hong Kong trip with their wife and 20-month-old. The columnist says activity planning took only about 30 minutes; Gemini drafted family-friendly daily plans that accounted for jet lag, to-do and packing lists, and found a cheap data plan. But the packing list left out the columnist's own socks and underwear, and in Hong Kong it suggested restaurants near the previous hotel in Taiwan; Ask Maps handled real-time requests better.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://cnaluxury.channelnewsasia.com/experiences/ai-travel-plan-gemini-vacation-293626 (2026-05-10)
- Last verified: 2026-10-17

### Ask ChatGPT whether a Schengen visa biometrics waiver applies, and miss the renewed-passport catch
- URL: https://www.realagentcases.com/use-cases/ask-chatgpt-whether-a-schengen-visa-biometrics-waiver-applies-and-miss-the-renew/
- Outcome: A client applying for an Austrian Schengen visa asked ChatGPT whether biometrics were mandatory. It said they can be waived if you've had a Schengen visa in the last five years, allowing a third party to submit. Visa professional Chandan Pandey noticed the client's previous visa was in an old passport, and the waiver requires it to be in the current one. Pandey refused to submit it on the client's behalf; the client went for biometrics in person and confirmed the rule. Pandey says relying on ChatGPT's answer would have meant a rejected application.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. ChatGPT gave the general rule but couldn't know the previous visa was in an old passport; following it would likely have led to a rejected application.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Reported by the visa professional handling the case, who has an interest in showing where AI falls short.
- Source: https://www.linkedin.com/posts/chandan-pandey-0836561b4_chatgpt-vs-visa-professional-client-applied-activity-7446074075816570881-w8Su (2026-04-04)
- Last verified: 2026-10-17

### Plan a vegetarian trip around Vietnam with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-vegetarian-trip-around-vietnam-with-gemini/
- Outcome: Over two weeks in Vietnam, Android Police's Parth Shah used Gemini to check weather by region, which showed he'd need warm clothes for Ba Na Hills and Sa Pa, and to build packing lists, shopping hotspots and a list of pure vegetarian restaurants he saved in Google Keep. On the trip it helped him check snacks for eggs and hidden allergens and explained the history behind temples and old towns.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.androidpolice.com/used-gemini-as-personal-translator-tour-guide-through-international-trip/ (2026-04-01)
- Last verified: 2026-10-17

### Win an airline's reimbursement appeal for a cancellation hotel and meals with Grok's step-by-step help
- URL: https://www.realagentcases.com/use-cases/win-an-airline-s-reimbursement-appeal-for-a-cancellation-hotel-and-meals-with-gr/
- Outcome: After a crew time-out cancelled their Miami flight, Chris Bernasconi was told by a gate agent that passengers were entitled to a hotel, transport and a meal, and saved every receipt and the airline's email. Delta declined the reimbursement claim. Bernasconi asked Grok, which spelled out what to do, how to frame the appeal and what to upload, and offered further steps if it failed. Within an hour of appealing, Delta said payment was on the way.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/ChrisCberna/status/2035728746947510753 (2026-03-22)
- Last verified: 2026-10-17

### Get a full refund after a flight returned to its origin, using rules Gemini found
- URL: https://www.realagentcases.com/use-cases/get-a-full-refund-after-a-flight-returned-to-its-origin-using-rules-gemini-found/
- Outcome: After their diverted flight returned to the origin airport in Delhi, a traveller used Gemini as a crisis consultant: they asked for the specific DGCA rules for a full refund instead of a credit shell, compared night train and bus reliability in fog, and got the exact Metro route to the interstate bus station. They say they secured a refund using the clauses Gemini gave them and were on a sleeper bus within an hour.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/chaithanya-moodbidri-prabhu-397952bb_traveltech-gemini-promptengineering-activity-7416836403340414976-VMoa (2026-01-13)
- Last verified: 2026-10-17

### Plan a 4-day family road trip to South Dakota with five chatbots and use the best plan
- URL: https://www.realagentcases.com/use-cases/plan-a-4-day-family-road-trip-to-south-dakota-with-five-chatbots-and-use-the-bes/
- Outcome: A Tom's Guide writer asked five chatbots to plan a four-day family trip from near Minneapolis to the Black Hills and Badlands, with gas and food stops, meals, lodging and under 1,400 miles in a borrowed minivan. Gemini was best at routing; ChatGPT did worst, with a broken map link and a needless night in the Badlands; Copilot was minimal; Claude was good apart from the route. DeepSeek won by basing the family in Custer State Park and planning every meal, and he followed its plan exactly, calling it a blast.
- Agents: DeepSeek, Gemini, ChatGPT, Claude, Copilot
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/i-made-five-chatbots-plan-my-holiday-it-wasnt-chatgpt-or-gemini-that-gave-me-the-best-response (2025-11-09)
- Last verified: 2026-10-17

### Untangle codeshare delay compensation across three jurisdictions with ChatGPT
- URL: https://www.realagentcases.com/use-cases/untangle-codeshare-delay-compensation-across-three-jurisdictions-with-chatgpt/
- Outcome: When a delay turned a 2-hour layover into an overnight stay in Turkey, a traveller first used ChatGPT to translate his correspondence for a Turkish-only support portal, then got the runaround: two airlines, a codeshare, three jurisdictions, and compensation firms that turned him down. Once he explained the full situation, he says ChatGPT reasoned through the rules, told him which airline to chase and what to say, and that airline agreed to compensation; he says he got $500.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/kunalkandekar_the-end-of-the-rip-off-economy-activity-7391950735816515585-QDee (2025-11-05)
- Last verified: 2026-10-17

### Plan a Tunisia trip with ChatGPT vs Claude, then travel the combined itinerary
- URL: https://www.realagentcases.com/use-cases/plan-a-tunisia-trip-with-chatgpt-vs-claude-then-travel-the-combined-itinerary/
- Outcome: A writer and his partner had ChatGPT and Claude each plan one day in Tunis, Tozeur and Hammamet, then built their trip from both. ChatGPT's plan was ambitious and cinematic; its Tozeur day 'nearly flattened' them with a far longer drive than implied, and a sandstorm blocked the road to the film sets. Claude's was slower and added two stops they loved, the Great Mosque of Kairouan and El Jem amphitheatre. He says ChatGPT delivered better memories and Claude more comfort and culture.
- Agents: ChatGPT, Claude
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.businessinsider.com/used-chatgpt-plan-trip-my-partner-used-claude-2025-10 (2025-10-12)
- Last verified: 2026-10-17

### Plan a dog-friendly 7-day RV trip with short drives using ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-dog-friendly-7-day-rv-trip-with-short-drives-using-chatgpt/
- Outcome: Planning an RV trip from Tampa with her partner and their dog, a poster asked ChatGPT for a 7-day route with no drive over 4 hours, waterfalls and short hikes, and dog-friendly stops. She also used it to check which state parks allow dogs on trails, how to connect a 30-amp RV to a 50-amp site, campfire cooking steps and campgrounds with cell service. She says it gave a day-by-day route with mileage and drive times, and dinner was saved.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.linkedin.com/posts/sstaartjes_just-came-back-from-an-amazing-inland-activity-7365782209313529858-bMSf (2025-08-25)
- Last verified: 2026-10-17

### Let a teenager run a California road trip itinerary with Grok and update it daily from family feedback
- URL: https://www.realagentcases.com/use-cases/let-a-teenager-run-a-california-road-trip-itinerary-with-grok-and-update-it-dail/
- Outcome: On a week-long drive north from Los Angeles on Highway 101, David Stevenson handed the itinerary to their daughter, who produced one with Grok within minutes and ran it through CarPlay. As the family spent longer at places they liked, she kept Grok updated and it revised the plan, adapting each day to their feedback. They visited places and restaurants they'd never have found themselves, all of which were great, with manageable driving between stops.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/davidgestevenson_been-a-fun-and-tiring-week-in-california-activity-7362686558937702402-0R-l (2025-08-17)
- Last verified: 2026-10-17

### Book a family member's return flight with ChatGPT agent, stopping before payment
- URL: https://www.realagentcases.com/use-cases/book-a-family-member-s-return-flight-with-chatgpt-agent-stopping-before-payment/
- Outcome: A poster asked ChatGPT agent to book a return flight for his mother-in-law with her specific seating and baggage preferences. He says it found the right flights, followed her instructions, clicked through the FlySafair website and stopped right before payment as asked; refining the prompt a few times made it more efficient.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/nicholas-lister-james-3a5606104_chatgpt-agent-just-booked-a-flight-for-me-activity-7356271644375416832-TQQC (2025-07-30)
- Last verified: 2026-10-17

### Handle parking, road tolls and EV charging questions on a European road trip with ChatGPT and Grok
- URL: https://www.realagentcases.com/use-cases/handle-parking-road-tolls-and-ev-charging-questions-on-a-european-road-trip-with/
- Outcome: On a family road trip through Germany, Austria, Italy, Croatia and Slovenia, a poster used ChatGPT and Grok as an on-the-go assistant: planning village visits, walks and beaches, finding fairly-priced safe parking, choosing local beers and wines, and answering first-time EV road-trip questions such as where to buy national vignettes online, road taxes and trip-time estimates.
- Agents: ChatGPT, Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/perferdinandsen_here-are-some-ways-i-have-used-chatgpt-activity-7355462062086459393-8_21 (2025-07-28)
- Last verified: 2026-10-17

### Make a personalized vacation packing checklist with a shared ChatGPT prompt
- URL: https://www.realagentcases.com/use-cases/make-a-personalized-vacation-packing-checklist-with-a-shared-chatgpt-prompt/
- Outcome: TechRadar's John-Anthony Disotto shares a long 'travel assistant' prompt (originally posted on Reddit) that asks for your age, destination and trip details, then builds a categorized packing checklist with clothing, tech, toiletries, luggage type and one commonly forgotten item. Trying it for his own trip, he says ChatGPT broke down his wardrobe into specific business-casual outfits, and that results are only as good as the details you give it.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task. The writer also ran it once for his own trip.
- Source: https://www.techradar.com/computing/artificial-intelligence/i-asked-chatgpt-to-help-me-pack-for-my-vacation-try-this-awesome-ai-prompt-that-makes-planning-your-travel-checklist-stress-free (2025-07-07)
- Last verified: 2026-10-17

### Plan an urgent passport agency visit before a vacation with ChatGPT
- URL: https://www.realagentcases.com/use-cases/get-an-expedited-passport-before-a-vacation-by-planning-a-passport-agency-visit/
- Outcome: With a passport expired just before a vacation, Dan Sanchez asked ChatGPT's o3 model to search for options. Together they worked out that Sanchez would need to drive to Atlanta for an expedited passport; when the government sites weren't clear about the process, ChatGPT filled gaps from Reddit and travel forums and helped with the appointment, hotels, where to be, what to bring, what to expect and backup plans. Sanchez says they got through the passport agency smoothly; the post doesn't say whether the passport was issued that day.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/digitalmarketingdan_danchez-activity-7348129028249149440-37zg (2025-07-07)
- Last verified: 2026-10-17

### Pick a holiday destination for three friends from their questionnaires with ChatGPT and Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/pick-a-holiday-destination-for-three-friends-from-their-questionnaires-with-chat/
- Outcome: For an AI-planned trip with two colleagues, TechRadar's Josephine Watson sent the same Deep Research prompt and three completed preference questionnaires to ChatGPT and Gemini: four days, one base, about $2,000 each, scenic and filming-friendly. Gemini returned five well-known picks such as Lagos and Mallorca in under five minutes. ChatGPT asked follow-up questions first, then ran 110 searches across 34 sources in 16 minutes and suggested Santorini, Madeira, the Bay of Kotor, Cappadocia and Essaouira. ChatGPT came out on top in their head-to-head.
- Agents: ChatGPT, Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.techradar.com/computing/artificial-intelligence/i-actually-followed-chatgpts-travel-advice-for-my-vacation-and-now-i-cant-hate-it-anymore (2025-06-19)
- Last verified: 2026-10-17

### Turn booked appointments into a realistic Washington DC itinerary with restaurants using Grok
- URL: https://www.realagentcases.com/use-cases/turn-booked-appointments-into-a-realistic-washington-dc-itinerary-with-restauran/
- Outcome: Ahead of a trip to DC, a poster gave Grok the appointments already set, the hotel locations and the remaining sights the group wanted to see. The poster says it produced an itinerary and, when prompted again, restaurants near the must-do stops to save searching once there, and called the result incredible.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/JamieLSaunders/status/1925972800700879008 (2025-05-23)
- Last verified: 2026-10-17

### Plan a month-long family road trip with kosher meals using ChatGPT and Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-month-long-family-road-trip-with-kosher-meals-using-chatgpt-and-grok/
- Outcome: Planning a month-long road trip for a family of seven, a poster used ChatGPT and Grok to map routes between national parks without backtracking, break long drives into kid-friendly segments, find lodging that fits, research kosher food and the kitchen kit needed for Shabbos, and create packing lists. He notes it didn't understand kosher certification symbols and that it's not perfect.
- Agents: ChatGPT, Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Written before the trip; the result isn't reported.
- Source: https://www.linkedin.com/posts/yitzy-feiglin-b9617867_familytravel-roadtripplanning-aiinreallife-activity-7327382891254943744-H0EP (2025-05-11)
- Last verified: 2026-10-17

### Get an Expedia cancellation refund with Grok's help reaching the right agent
- URL: https://www.realagentcases.com/use-cases/get-an-expedia-cancellation-refund-with-grok-s-help-reaching-the-right-agent/
- Outcome: Stuck in Expedia's cancellation process and not getting the refund he believed he was owed, a poster asked Grok for help and found the Expedia phone number. He says the Expedia agent then helped him get a refund.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.linkedin.com/posts/ryanjohnrussell_grok-activity-7321940173502431234-58Ju (2025-04-26)
- Last verified: 2026-10-17

### Find the airline's hidden refund option instead of a voucher with Gemini
- URL: https://www.realagentcases.com/use-cases/find-the-airline-s-hidden-refund-option-instead-of-a-voucher-with-gemini/
- Outcome: After a day of 4 delays, 4 rebookings and a last-minute diversion, a traveller was handed a Southwest form that, if filled in, gave up the right to a refund in exchange for a voucher. He asked Gemini whether there was really no refund option, and it gave him a near-identical link that included a refund request.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: The poster doesn't say whether the refund came through.
- Source: https://www.linkedin.com/posts/zachrattner_i-discovered-a-dark-ux-pattern-today-expectation-activity-7280416699646988288-vUxW (2025-01-02)
- Last verified: 2026-10-17

### Plan and book a trip with Grok Bot (7-step prompt set)
- URL: https://www.realagentcases.com/use-cases/plan-and-book-a-trip-with-grok-bot-7-step-prompt-set/
- Outcome: A shared sequence of seven prompts for planning a trip with Grok Bot: a travel brief, three priced options, a fare watch, a booking step with a confirm-before-paying check, a paperwork checklist, a day-by-day itinerary and a day-before briefing. The post is a prompt guide; it doesn't show a completed booking or any results.
- Agents: Grok Bot
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: Prompt guide: a shared set of prompts, not a record of a completed trip.
- Disclosure: This is our product: realagentcases is run with Grok Bot. Included because the source is public and verifiable.
- Source: https://x.com/free_ai_guides/status/2111004261488480425 (2026-10-16)
- Last verified: 2026-10-16

### Find flights with Manus (16-hour run, poor pick)
- URL: https://www.realagentcases.com/use-cases/find-flights-with-manus-16-hour-run-poor-pick/
- Outcome: Asked to find flights, Manus ran for 16 hours and then picked what the poster called probably the most expensive option, with a 9-hour layover.
- Agents: Manus
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Ran for 16 hours, then picked what the poster called probably the most expensive option, with a 9-hour layover.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/natalienkhalil/status/2107519229109182513 (2026-10-06)
- Last verified: 2026-10-16

### Plan a six-week Southeast Asia trip with Manus (hallucinated flights)
- URL: https://www.realagentcases.com/use-cases/plan-a-six-week-southeast-asia-trip-with-manus-hallucinated-flights/
- Outcome: The poster used Manus to plan a 6-week Southeast Asia trip, and it hallucinated flights that didn't exist despite having browser access.
- Agents: Manus
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Recommended flights that did not exist, despite having browser access.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/MattMickiewicz/status/2103222377459188214 (2026-09-24)
- Last verified: 2026-10-16

### Book every restaurant for a Japan trip with Muse
- URL: https://www.realagentcases.com/use-cases/book-every-restaurant-for-a-japan-trip-with-muse/
- Outcome: Muse booked every restaurant for the poster's Japan trip and corresponded in Japanese with a Hakone hotel to secure a dinner slot not offered on its website.
- Agents: Muse
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/Han_Fang_/status/2101874865385169348 (2026-09-21)
- Last verified: 2026-10-16

### Book a whole weekend trip in 30 minutes with ChatGPT Work
- URL: https://www.realagentcases.com/use-cases/book-a-whole-weekend-trip-in-30-minutes-with-chatgpt-work/
- Outcome: ChatGPT read the poster's chat with their girlfriend, proposed itinerary options, staged an Airbnb, rental car, fishing gear and dinner reservations through browser use, then made a shareable trip page after the poster booked.
- Agents: ChatGPT Work
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/nickbaumann_/status/2087598230968373584 (2026-08-12)
- Last verified: 2026-10-16

### Plan a food-focused trip on a fixed budget with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-food-focused-trip-on-a-fixed-budget-with-gemini/
- Outcome: Android Police's Akshay Bhalla gave Gemini a Rs 1,00,000 budget for a September trip to North Goa with flights and stays already booked, and a cap of Rs 3,000-4,000 per meal. He says Gemini used Google Maps to suggest restaurants and activities in his favourite neighbourhoods, mapped routes and best visiting times, and found hidden gems and beach shacks after he refined the prompt.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.androidpolice.com/gemini-vacation-planning-budget/ (2026-08-09)
- Last verified: 2026-10-16

### Plan a solo summer vacation with Gemini (forgot it was a solo trip)
- URL: https://www.realagentcases.com/use-cases/plan-a-solo-summer-vacation-with-gemini-forgot-it-was-a-solo-trip/
- Outcome: XDA's Mahnoor Faisal let Claude, ChatGPT and Gemini plan a real summer vacation she was taking alone. She says Gemini kept forgetting basics, repeatedly defaulting to group and couple-oriented ideas like a romantic candlelight dinner, and tended to agree her plan was great when it needed work.
- Agents: Gemini
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Gemini kept defaulting to group and couple suggestions and tended to agree with an over-packed plan.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.xda-developers.com/let-claude-chatgpt-gemini-plan-my-vacation/ (2026-08-09)
- Last verified: 2026-10-16

### Claim compensation for a long flight delay (EU261) with ChatGPT
- URL: https://www.realagentcases.com/use-cases/claim-compensation-for-a-long-flight-delay-eu261-with-chatgpt/
- Outcome: After a Norse Atlantic flight from Oslo was delayed overnight and the airline offered a $25 refreshment card, then ignored his claims, the author pasted the delay email into ChatGPT Pro and asked about his rights. Over eleven months it walked him through EU261 (600 euros per passenger for four people), the national complaint body, the Norwegian conciliation court and a settlement including statutory interest; on July 14, 2026 the airline wired $4,760.36.
- Agents: ChatGPT
- Category: Travel
- Evidence: Described by poster
- Claimed result: $4,760.36 paid after eleven months (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.behind-the-enemy-lines.com/2026/07/the-lawyer-i-never-hired-how-chatgpt.html (2026-07-22)
- Last verified: 2026-10-16

### Get restaurant and transport tips on a Hong Kong family trip with Grok
- URL: https://www.realagentcases.com/use-cases/get-restaurant-and-transport-tips-on-a-hong-kong-family-trip-with-grok/
- Outcome: On a 6-day family trip to Hong Kong and Macau, the PickEnough author checked AI travel advice against reality. ChatGPT and Claude were blocked by regional restrictions; of the tools that worked (Grok, Gemini and Perplexity), the author rated Grok the most accurate overall (7.2/10), with correct real-time navigation and local restaurant picks. Asked what was good to eat within 5 minutes while walking the Mid-Levels, Grok suggested a nearby egg-tart shop the author rated 4/5. Across the tools, transport predictions were within 10%, but restaurant cost estimates assumed 2 people even though the group was 4.
- Agents: Grok
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://pickenough.com/chatgpt-gemini-grok-travel-test/ (2026-06-13)
- Last verified: 2026-10-16

### Build a work-trip itinerary from booking emails with Claude
- URL: https://www.realagentcases.com/use-cases/build-a-work-trip-itinerary-from-booking-emails-with-claude/
- Outcome: The day before a work trip, Claude read the poster's booking emails and organized a polished itinerary that the poster says would have taken a few hours by hand.
- Agents: Claude
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/andreistenmark/status/2064466188764418315 (2026-06-09)
- Last verified: 2026-10-16

### Compare Lisbon hotels across three date ranges with Gemini in Chrome
- URL: https://www.realagentcases.com/use-cases/compare-lisbon-hotels-across-three-date-ranges-with-gemini-in-chrome/
- Outcome: Asked for a central Lisbon hotel for two under $220 a night near Alfama across three October week-long windows, Chrome auto browse opened Booking, Hotels.com, Expedia and Hyatt, kept a running comparison and surfaced three candidates with trade-offs in about six minutes, stopping before booking.
- Agents: Gemini in Chrome (auto browse)
- Category: Travel
- Evidence: Article or interview
- Claimed result: About six minutes (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://blog.imseankim.com/chrome-auto-browse-gemini-agentic-hotel-form-fill-google-io-2026/ (2026-05-20)
- Last verified: 2026-10-16

### Plan a summer road trip by answering questions one at a time with ChatGPT
- URL: https://www.realagentcases.com/use-cases/plan-a-summer-road-trip-by-answering-questions-one-at-a-time-with-chatgpt/
- Outcome: In a Tom's Guide road-trip feature, Elton Jones shares an all-in-one prompt (written with ChatGPT's help) that has the chatbot ask about starting point, budget, trip length, travel style, food, music, routes, lodging and companions, then build an itinerary, driving schedule, budget, packing list, food stops, scenic stops and rain plans. It offers prompts only and doesn't report a finished trip.
- Agents: ChatGPT
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: Prompt guide: a shared prompt, not a record of a completed task.
- Source: https://www.tomsguide.com/ai/7-easy-ways-to-use-ai-to-plan-your-entire-2026-road-trip-in-minutes (2026-05-17)
- Last verified: 2026-10-16

### Plan a family road trip with playground stops and a dog-friendly hotel with Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-family-road-trip-with-playground-stops-and-a-dog-friendly-hotel-with-grok/
- Outcome: Jeff Binek, a gym owner in Dublin, Ohio, asked Grok to plan a two-day drive from Columbus to Hilton Head via Charlotte with three kids under four and a dog, stopping every 2.5 hours at safe playgrounds near gas stations. He says it gave precise timing and distances, gas prices, three park options per stop and a dog-friendly hotel, and the family reversed the same plan on the way home.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://newsletter.cbusaiagents.com/p/ai-agents-explained-columbus-small-business (2026-03-31)
- Last verified: 2026-10-16

### Shortlist cheap, quiet places to stay for a holiday weekend with Gemini
- URL: https://www.realagentcases.com/use-cases/shortlist-cheap-quiet-places-to-stay-for-a-holiday-weekend-with-gemini/
- Outcome: For a New Year's trip to Oxford, MakeUseOf's Jack Mitchell asked Gemini for 5 hotels or B&Bs in or around the city that were cheapest and not overcrowded over the holiday. He says Gemini shortlisted places factoring in seasonal crowds and typical costs, and he then compared travel times, tolls and fuel or transit costs in Google Maps before choosing.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.makeuseof.com/ways-use-google-maps-gemini-to-plan-cheaper-less-stressful-weekend-trip/ (2026-02-16)
- Last verified: 2026-10-16

### Plan a vacation with flight, hotel and busy-week checks in Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-vacation-with-flight-hotel-and-busy-week-checks-in-gemini/
- Outcome: A Hacker News commenter says they've been using Gemini to plan a vacation and finds it surprisingly useful: it can look up flights from Google Flights and hotels from Maps, and knows about busy weeks at the destination. They add that you have to hope it isn't hallucinating, which they see as safer for trips than for healthcare.
- Agents: Gemini
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://news.ycombinator.com/item?id=46825203 (2026-01-30)
- Last verified: 2026-10-16

### Plan a family Disneyland trip with Claude Code and Skills
- URL: https://www.realagentcases.com/use-cases/plan-a-family-disneyland-trip-with-claude-code-and-skills/
- Outcome: The poster planned a family Disneyland trip using Claude Code as the agent, Claude Skills for research, planning, tracking and journaling, and a simple companion web app for itineraries and maps.
- Agents: Claude Code
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/gautam/status/2012601709000814926 (2026-01-17)
- Last verified: 2026-10-16

### Plan a five-day city trip with flights in Google AI Mode's Canvas
- URL: https://www.realagentcases.com/use-cases/plan-a-five-day-city-trip-with-flights-in-google-ai-mode-s-canvas/
- Outcome: ZDNet's tester asked Google's AI Mode trip-planning tool to plan a five-day New York City trip focused on historical sites and museums, then asked it to add flights; it researched Charlotte-to-NYC options and added a Flights tab with three major-airline choices to the itinerary. The tester says everything was well organized in Canvas, with Google reviews alongside each stop.
- Agents: Google AI Mode
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.zdnet.com/article/i-tried-googles-new-trip-planning-ai-tool-and-ill-never-plan-my-own-trip-again/ (2025-11-18)
- Last verified: 2026-10-16

### Find nonstop flights and a hotel under $400 a night with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-nonstop-flights-and-a-hotel-under-400-a-night-with-chatgpt-agent/
- Outcome: BKA Content's Greg Secrist asked ChatGPT agent for nonstop Salt Lake City to Las Vegas flights for October 8–11 plus hotel rooms under $400 per night. It worked across Delta, Expedia and Booking.com but struggled with JavaScript-heavy pages and date pickers, sometimes looping or freezing; after almost nine minutes it returned some options he says he could have found manually in half the time.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Struggled with JavaScript-heavy pages and date pickers, looping or freezing; results took almost nine minutes.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://bkacontent.com/how-to-use-chatgpt-agent-mode/ (2025-10-29)
- Last verified: 2026-10-16

### Book a car rental with ChatGPT Atlas agent mode
- URL: https://www.realagentcases.com/use-cases/book-a-car-rental-with-chatgpt-atlas-agent-mode/
- Outcome: The poster asked ChatGPT Atlas in agent mode to book a car rental and says it did, shown in a short demo video.
- Agents: ChatGPT Atlas agent mode
- Category: Travel
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/upendra_dama/status/1981677020728656302 (2025-10-24)
- Last verified: 2026-10-16

### Save 20 restaurants from an article to a Google Maps list with Atlas
- URL: https://www.realagentcases.com/use-cases/save-20-restaurants-from-an-article-to-a-google-maps-list-with-atlas/
- Outcome: ChatGPT Atlas agent picked 20 restaurants from an article and added them to a new saved-places list in the poster's Google Maps account; the poster called it slow but an excellent result.
- Agents: ChatGPT Atlas agent mode
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/GershomCharig/status/1980747058110943580 (2025-10-21)
- Last verified: 2026-10-16

### Add flights from a confirmation email to Google Calendar with Comet
- URL: https://www.realagentcases.com/use-cases/add-flights-from-a-confirmation-email-to-google-calendar-with-comet/
- Outcome: From a flight confirmation email, Comet Assistant opened Google Calendar and created entries with airports and flight numbers for a NJ–London–Dubai–NJ trip, but the entries didn't span the full flight times and the poster had to fix them by hand and re-prompt for the check-in reminders.
- Agents: Perplexity Comet Assistant
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Flight events had the wrong lengths and the reminders needed a second prompt.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/mikewchan/status/1970911237438415050 (2025-09-24)
- Last verified: 2026-10-16

### Find the cheapest Hamburg–Copenhagen train with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-the-cheapest-hamburg-copenhagen-train-with-chatgpt-agent/
- Outcome: Asked (in German) to book the cheapest Hamburg–Copenhagen train for September, ChatGPT agent found a €63.26 fare after about ten minutes on the Deutsche Bahn site, then asked the author to log in; Comet was faster (3:40) but booked legs separately for €70.94.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Article or interview
- Claimed result: €63.26 vs €70.94 from Comet (Poster says; a claim from the source, not independently verified)
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://olereissmann.com/i-tried-chatgpt-agent-and-perplexity-comet-to-actually-get-things-done/ (2025-08-04)
- Last verified: 2026-10-16

### Plan a points-funded family holiday and book dinner with ChatGPT agent (partial)
- URL: https://www.realagentcases.com/use-cases/plan-a-points-funded-family-holiday-and-book-dinner-with-chatgpt-agent-partial/
- Outcome: For a points-funded family holiday plan, ChatGPT agent built a decent deck but spent 15 minutes formatting slides; it couldn't book the romantic dinner.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Built a decent deck but spent 15 minutes formatting slides and could not book the dinner.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/hotelemarketer/status/1949081438076244273 (2025-07-26)
- Last verified: 2026-10-16

### Check a hotel price with ChatGPT agent (wrong town)
- URL: https://www.realagentcases.com/use-cases/check-a-hotel-price-with-chatgpt-agent-wrong-town/
- Outcome: Asked to check a hotel's price with its town clearly stated, ChatGPT agent looked up a same-named hotel in another town, and the poster notes this wouldn't have been obvious from the reply.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Looked up a same-named hotel in the wrong town, and the reply did not make that obvious.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/szilagyipal/status/1948670475959501318 (2025-07-25)
- Last verified: 2026-10-16

### Build a saved Google Maps list of a city's best restaurants with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/build-a-saved-google-maps-list-of-a-city-s-best-restaurants-with-chatgpt-agent/
- Outcome: Frank Andrade asked ChatGPT agent to find the best restaurants in the city he was about to visit and save them as a new Google Maps list with notes on signature dishes. It picked 10 restaurants, collected their signature dishes, paused for him to enter his Google login, then created and saved the list with notes; it took about 23 minutes.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Shown on video
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://artificialcorner.com/p/my-honest-review-of-chatgpt-agent (2025-07-18)
- Last verified: 2026-10-16

### Compare 7-day Bali itineraries from ChatGPT and Grok
- URL: https://www.realagentcases.com/use-cases/compare-7-day-bali-itineraries-from-chatgpt-and-grok/
- Outcome: An X user asked both Grok and ChatGPT for a 7-day solo travel itinerary to Bali and says ChatGPT did a better job, with tables and accurate breakdowns of the expenses the poster expected.
- Agents: ChatGPT, Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Context: In a comparison test: the poster's own side-by-side of two chatbots, not a standalone agent test.
- Source: https://x.com/Invincible39/status/1910574756505477364 (2025-04-11)
- Last verified: 2026-10-16

### Plan a toddler-friendly San Diego itinerary with Gemini Deep Research
- URL: https://www.realagentcases.com/use-cases/plan-a-toddler-friendly-san-diego-itinerary-with-gemini-deep-research/
- Outcome: Gemini Deep Research produced a San Diego spring-break itinerary covering the zoo, SeaWorld and Balboa Park built around a daily nap for a two-year-old; the poster called the report insightful and detailed.
- Agents: Gemini Deep Research
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/maizeandhsu/status/1909832845222461680 (2025-04-09)
- Last verified: 2026-10-16

### Plan a travel itinerary with map-ready stops, food and restroom finds with Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-travel-itinerary-with-map-ready-stops-food-and-restroom-finds-with-grok/
- Outcome: An X user recommends Grok for planning travel itineraries, saying it can create a timeline, give written driving locations to plug into any map app, offer restaurant suggestions and find public toilet locations, and that it created an itinerary the poster calls amazing when asked what to do.
- Agents: Grok
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/tidbits84/status/1904946476956131741 (2025-03-26)
- Last verified: 2026-10-16

### Rank safe cities for solo travel with Grok DeepSearch
- URL: https://www.realagentcases.com/use-cases/rank-safe-cities-for-solo-travel-with-grok-deepsearch/
- Outcome: Asked for the safest destinations for solo travelers in 2025, Grok DeepSearch returned a table of cities with Global Peace Index scores, places to visit and solo-travel tips; TechWiser's reviewer said it stood out for structure and focus on solo travel.
- Agents: Grok DeepSearch
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://techwiser.com/groks-deepsearch-vs-chatgpts-deepresearch-we-tested-with-5-prompts/ (2025-03-04)
- Last verified: 2026-10-16

### Plan a 3-day Tokyo tech trip with a budget in USD and yen with Grok
- URL: https://www.realagentcases.com/use-cases/plan-a-3-day-tokyo-tech-trip-with-a-budget-in-usd-and-yen-with-grok/
- Outcome: In Tom's Guide's ChatGPT-vs-Grok test, Ryan Morrison asked both for a 3-day Tokyo technology trip with hotels, transport, a robot restaurant and a budget in USD and yen; he judged Grok's itinerary more focused, realistic and detailed, with a fuller budget breakdown.
- Agents: Grok
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Context: In a comparison test: from a head-to-head chatbot review, not a standalone agent test.
- Source: https://www.tomsguide.com/ai/chatgpt/i-put-chatgpt-vs-grok-to-the-test-with-7-prompts-heres-the-winner (2025-01-09)
- Last verified: 2026-10-16

### Plan a 10-day Malaysia trip with flights and hotels with Gemini
- URL: https://www.realagentcases.com/use-cases/plan-a-10-day-malaysia-trip-with-flights-and-hotels-with-gemini/
- Outcome: Hulry's writer asked Gemini to plan a 10-day trip from Bangalore to Malaysia for two at a moderate pace, low-to-mid budget with hotels under ₹5,000 a night. Gemini returned a decent itinerary and, using Google Flights data, flight options with accurate timings and pricing, but it mixed up the travel dates while searching for flights. The article says the whole plan took under 15 minutes.
- Agents: Gemini
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Partial. Gemini mixed up the travel dates when searching for flights.
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://hulry.com/google-ai-travel-planner/ (2024-03-27)
- Last verified: 2026-10-16

### File Thailand Digital Arrival Cards for five travelers with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/file-thailand-digital-arrival-cards-for-five-travelers-with-chatgpt-agent/
- Outcome: ChatGPT agent filled Thailand Digital Arrival Cards for 5 people (about an hour / 16 credits); the poster verified the preview and swapped in real passport numbers before submitting.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/SatanAtWink/status/2006301585996169399 (2025-12-31)
- Last verified: 2026-10-16

### Hunt last-minute holiday flights with multiple agents (failure case)
- URL: https://www.realagentcases.com/use-cases/hunt-last-minute-holiday-flights-with-multiple-agents-failure-case/
- Outcome: Claude Opus 4.5, Gemini 3 Pro and ChatGPT agent mode only matched a manual Google Flights search and were too slow as cheap fares vanished; the humans beat them by about 15%.
- Agents: ChatGPT agent, Claude Opus 4.5 (Antigravity), Gemini 3 Pro
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. The agents only matched a manual Google Flights search and were too slow as cheap fares vanished; the humans beat them by about 15%.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/anuraagsidhu/status/2001930515260088374 (2025-12-19)
- Last verified: 2026-10-16

### Compare browser agents on a vacation Airbnb search
- URL: https://www.realagentcases.com/use-cases/compare-browser-agents-on-a-vacation-airbnb-search/
- Outcome: On the same Airbnb search, ChatGPT Atlas agent mode and Comet each found properties in about 5 minutes, Claude's extension took 10 minutes to find the first, and MiniMax Agent ran over an hour without giving links.
- Agents: ChatGPT Atlas agent mode, Perplexity Comet, Claude in Chrome, MiniMax Agent
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/oyacaro/status/1982805076952465599 (2025-10-27)
- Last verified: 2026-10-16

### Plan St. Simons Island restaurants with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/plan-st-simons-island-restaurants-with-chatgpt-agent/
- Outcome: ChatGPT Agent Mode produced a dining guide for a 4-day St. Simons Island trip organized by meal and price, and the author ate at 4 restaurants from the list and liked them all.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Article or interview
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://www.theaienterprise.io/p/ai-agents-research (2025-10-21)
- Last verified: 2026-10-16

### Search for cheap flights with ChatGPT agent (failure case)
- URL: https://www.realagentcases.com/use-cases/search-for-cheap-flights-with-chatgpt-agent-failure-case/
- Outcome: ChatGPT agent suggested flights costing $1,400 when the poster found an $850 option on their own.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Failure. Suggested a $1,400 flight; the poster found an $850 option on their own.
- Prompt type: Improved (strengthened by us to avoid the problem in the source; not tested by the poster). The prompt they actually used: recreated by us
- Source: https://x.com/maxdrescher_/status/1958253813770674556 (2025-08-20)
- Last verified: 2026-10-16

### Find the cheapest one-way flight across two departure cities with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-the-cheapest-one-way-flight-across-two-departure-cities-with-chatgpt-agent/
- Outcome: From a vague brief, ChatGPT agent searched flights from Ljubljana or Zagreb to Bordeaux over a two-week window and found a €141.48 fare in 12 minutes.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Screenshots or logs
- Claimed result: None
- Beginner-friendly: Yes
- Outcome status: Worked (per the source)
- Prompt type: Original (as published in the source)
- Source: https://x.com/GasperCrepinsek/status/1951266026324205864 (2025-08-01)
- Last verified: 2026-10-16

### Plan a Budapest weekend and start an Airbnb booking with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/plan-a-budapest-weekend-and-start-an-airbnb-booking-with-chatgpt-agent/
- Outcome: ChatGPT agent researched accommodation, activities and vegan restaurants for a Budapest weekend and got into an Airbnb reservation flow, then handed over to the user for login and payment.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://www.visibletourism.com/post/from-zero-click-to-zero-effort-chatgpt-agent-mode/ (2025-07-25)
- Last verified: 2026-10-16

### Book an Airbnb with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/book-an-airbnb-with-chatgpt-agent/
- Outcome: ChatGPT agent booked what the poster called the perfect Airbnb on the first try.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/romainhuet/status/1946015036503449813 (2025-07-18)
- Last verified: 2026-10-16

### Find a family vacation rental in Costa Rica with ChatGPT agent
- URL: https://www.realagentcases.com/use-cases/find-a-family-vacation-rental-in-costa-rica-with-chatgpt-agent/
- Outcome: ChatGPT agent found a Costa Rica place with a pool at a decent rate for 2 adults and 3 kids; the poster was tempted to book without checking further.
- Agents: ChatGPT agent
- Category: Travel
- Evidence: Described by poster
- Claimed result: None
- Beginner-friendly: No
- Outcome status: Worked (per the source)
- Prompt type: Recreated (the source did not publish the exact prompt)
- Source: https://x.com/kalenjordan/status/1946285767690727918 (2025-07-18)
- Last verified: 2026-10-16