Coordinate agents in Slack and Cursor to propose growth experiments for human review. Ezra’s Auto XP system was still being developed at the time of the episode.
with Claire Vo
One of the things you can do is just say, 'Hey, Auto XP, go find an experiment to run on our settings page. See what you can do. Go crazy.' And it will trigger a data scientist agent to go find drop-off and like look at the funnel... It will trigger a UX writing agent that will come up with some ideas for copy... Lastly, it'll trigger an engineer that'll pick up a PR and run it. And then again, with all of these steps, it writes back to the Slack channel in the same thread.
A documented experiment proposal and draft implementation for human review.
AI-first code editor
A workplace messaging and collaboration platform.
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
7 steps
Create a dedicated Slack channel (e.g., #auto-xp). Using a tool like Cursor, configure three distinct agent automations: a 'data scientist', a 'UX writer', and an 'engineer'. Each agent should be configured to trigger the next and write its output into the same Slack thread.
This setup creates a transparent, observable workflow where a human can follow the experiment's entire lifecycle in one place.
In the dedicated Slack channel, post a message to kick off the workflow. This can be a broad directive or a more targeted request.
"Hey, Auto XP, go find an experiment to run on our settings page. See what you can do. Go crazy."
This initial prompt triggers the first agent in the chain, the data scientist, to begin its analysis.
The data scientist agent analyzes product data (e.g., funnels, drop-off) to find an opportunity. It then formulates a hypothesis and posts its findings to the Slack thread.
Check that the agent’s proposed opportunity is supported by the underlying data.
Triggered by the data scientist's output, the UX writing agent (and potentially a design agent) proposes a specific change, such as new copy or a modified layout, and adds it to the thread.
This agent translates the data-driven opportunity into a concrete user-facing change.
The engineering agent takes the proposed solution and creates a pull request to implement the experiment. It posts a link to the PR in the Slack thread.
The PR keeps the implementation alongside the hypothesis for review.
A product manager reviews the entire Slack thread, which now contains the data, hypothesis, proposed solution, and code. If approved, the experiment is launched.
This is the most critical human checkpoint. The goal is to compress the work, not to abdicate responsibility for what gets shipped to users.
Configure a webhook to trigger an analysis agent when the experiment concludes. This agent should analyze the results and post a summary, then automatically kick off the next iteration based on the findings.
Review the analysis before choosing the next experiment; this loop was still being developed in the episode.
Turn an idea into a PRD, user stories, and a plan.
How to recover when the loop fails and where human judgment helps.
Let routine work keep moving.
Create a repository-based system to provide an AI assistant with durable context about your team, projects, and communication channels. This enables it to perform routine tasks like summarizing daily updates and identifying stakeholders, acting as a personalized chief of staff.
Use Meta's Muse to generate a printable, multi-page family newspaper that includes a weekly schedule, weather forecasts, and activity ideas. This workflow leverages Muse's strength in creating visually appealing, family-oriented artifacts.
Use Claude Duet to generate a complete product launch video that explains the value proposition, incorporates branding, and includes a synthetic voiceover from ElevenLabs. This workflow is ideal for creating polished, walkthrough-style product videos.
Join 250,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.
