Process a bounded set of LinkedIn requests with explicit acceptance rules and a reviewable action log.
Source episode · 33:46
How I AI: My GPT-5.6 Sol Benchmark & 4 Game-Changing Workflows (vs. Fable)with Claire Vo
33:46 to 35:47: Claire gives GPT-5.6 browser control a LinkedIn page and rules for which connection requests to accept, then watches it work through the interface.
A completed batch whose actions and skipped decisions are recorded for review.
OpenAI's cloud-based AI software engineering agent that can execute code, run tests, and handle complex multi-file tasks autonomously.
Step by step
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
5 steps
In your Chrome browser, open the page you want to automate and ensure you are logged in. For this workflow, navigate to your LinkedIn messages page.
Open the exact LinkedIn request queue and do not navigate to unrelated pages.
In your AI tool, such as Codex, type the command to activate browser control. For example, enter @Chrome to give the AI control of your active Chrome tab.
Enable browser control for this tab and confirm the requested batch limit.
Provide a detailed prompt that defines the task and includes strict, non negotiable criteria. For managing an inbox, specify what makes a message high value, such as the sender's job title or company tier, to ensure the AI only acts on the correct items.
For the next [number] requests: accept only when [visible rules], decline only when [rules], and skip anything ambiguous. Never send a message.
Execute the prompt and let the AI begin working in the browser. Monitor the AI's first few actions in real time to confirm it is correctly interpreting your rules before letting it run unsupervised.
Process one profile at a time, log name, visible evidence, decision, and action, then stop on any warning or challenge.
After the automation finishes, review the actions it took, such as messages sent and connections accepted. Verify their quality and accuracy to refine your prompt for next time. The author noted it successfully processed hundreds of messages.
Compare the final log with LinkedIn state and report accepted, declined, skipped, and untouched counts.
Turn an idea into a PRD, user stories, and a plan.
After the steps
How to recover when the loop fails and where human judgment helps.
Recover
Keep building
Use a hybrid AI workflow to bridge design and development. Generate high-level design concepts and tokens with Claude, then use Codex for robust implementation of the technical components in your codebase.
Implement a QA loop for AI-generated code using Devin. This playbook combines automated code reviews with an agent-recorded video walkthrough for visual verification, ensuring high-quality merges at scale.
Build a reusable 'Watchdog' playbook in Devin to automatically monitor customer accounts. The agent proactively checks for errors, synthesizes a prioritized list of critical issues, and saves you time.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.