Skip to main content
I want my LLM to understand the user flow of the product I’m working on (useful context for AI for brainstorming ideas, instrumentation, generating hypotheses, drafting artifacts… anything product related). TLDR: What worked best for me was: Loom walkthrough → upload to Gemini Pro → context for my PM AI copilot 🥳
  1. I did a 2-min Loom walkthrough of the specific flow I’m working on
I give a tour and explain this as if I’m doing a help center video or onboarding a new team member. Above is an example of me testing it out on Surfline—ahem, a data analytics platform. To be especially thorough, you could ask your local knowledge team for the video files they’ve already created to the help center.
  1. I uploaded the resulting video file (!) to Gemini Pro 2.5 with this prompt:
This is a video of me giving a tour of a surfing forecast web app, can you please outline the user flow in as much detail as possible optimized for AI to use as context. (Claude doesn’t take video. ChatGPT 4o and 5 took video files, but didn’t give useful results.)
  1. I save the result as a document, and add to project knowledge of my AI copilot (gemini gem, chatgpt/claude project, file in cursor, etc.—or perhaps an AI prototyping tool?)
Now my AI copilot has remarkably granular context on what I’m working on, from a user experience point of view. I put this in the project knowledge, since it’s the kind of context you can put in project knowledge and transcends any particular initiative, so it can help you whatever you’re working on.

​
Get notified about future posts

Subscribe

​
Let’s build together

If you’re reading this, you’re in the top 5% of people leaning into AI agents for real work (even if it doesn’t feel that way, trust me). To go deeper, join me for a live, hands-on workshop. To get your team on board, let’s plan an AI build sprint.