Test Operator on legal paperwork, party planning and shopping
The author tried Operator on five use cases; party planning was 'meh' and Amazon shopping failed because it couldn't add items to the cart.
Done withOperator
- What they did
- Cassie Kozyrkov used OpenAI's Operator, a browser-controlling agent available to ChatGPT Pro subscribers, and screen-recorded it on several tasks. The five featured in the video were hotels, data gathering, flights, legal paperwork and infographics. She also tried kids' party planning with web search, Amazon shopping and finding a freelancer, and left those three out of the video.
- How it went
- Party planning was 'meh'. Amazon shopping failed because Operator couldn't add items to the cart. The freelancer task went badly, which she blames on her own prompt. Operator did better when 'good enough' sufficed.
- Worth knowing
- Operator is very slow: about 2.5 hours of recording became 18 minutes of video. It needs a US-based $200/month Pro plan, and it can't run tasks in parallel or in a loop.
Try it yourself with Operator
Help me with this errand: [e.g. plan a birthday party for [number] guests on [date] / find and add [items] to my cart on [store site]]. Work step by step and report what worked and what got stuck. Stop and show me the cart or plan before you pay, submit any form, or send anything, and give me a short list of what you finished and what I need to do myself.
Source: decision.substack.com · Undated
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Find a toy and start checkout with Operator
OpenAI's Operator picked a toy matching criteria and opened Target.com but did not complete the purchase, walking the user through checkout.
Find London tours and order groceries
Operator found top-rated London walking tours on TripAdvisor, but its Instacart order went wrong by searching in the wrong city.
Research sales leads with Operator
Testers ran Operator on lead research; after a CAPTCHA handover it lost its mission and returned generic company profiles instead of qualified leads.
Get a second opinion on an MRI
One person's account of using Claude Code to analyze their own MRI.
Clear leftover installers and fetch a YouTube transcript
Scout ran PowerShell to delete 31 leftover installers (4.4 GB) after asking permission, and installed a tool to save a YouTube transcript when the browser was blocked.
Assemble an immigration document package
A user had ChatGPT browser use gather seven years of taxes, bank statements and immigration documents into a package in minutes.