Stop Losing Your Best Ideas: How Product Managers Are Using ChatGPT as a 'Second Brain'

The Problem

Imagine you're a product manager preparing for a big strategy meeting. You know your team discussed something relevant six months ago—maybe it was user feedback about a feature, or lessons from a failed launch. But where did you put those notes? Was it in Notion? A Google Doc? An email thread?

This is the daily reality for most product managers. You accumulate incredibly valuable knowledge through your work—insights from user interviews, competitive analyses, post-mortems from past projects—but it all ends up scattered across dozens of tools and folders. When you actually need that information, it's either impossible to find or you've forgotten it exists entirely.

The result? You end up reinventing the wheel, making the same mistakes twice, or going into important meetings less prepared than you could be.

The Solution Explained

The 'second brain' concept comes from productivity expert Tiago Forte, but Lenny Rachitsky has adapted it specifically for product managers using ChatGPT. Think of it as building a personal research assistant that remembers everything you've ever learned on the job.

The system works in three simple phases:

The magic is that ChatGPT does the heavy lifting of organizing your messy notes into clean, searchable 'knowledge cards'—each one containing a summary, key insights, action items, and relevant tags.

How It Actually Works

Step 1: Turn Chaos Into Structure

Let's say you just finished a user interview with messy, stream-of-consciousness notes. You paste them into ChatGPT along with a prompt like:

"Summarize this user interview into a knowledge card with: a 2-sentence summary, 3 key insights, any action items, and suggested tags."

In under 60 seconds, ChatGPT transforms your rambling notes into a clean, structured card you can actually use later.

Step 2: Build Your Searchable Library

These knowledge cards go into Notion, where you create a simple database. Each card gets properties like 'Topic,' 'Date,' 'Source Type' (interview, meeting, research), and 'Project.' Over time, you build a personal library of everything you've learned.

The real power emerges when patterns start appearing. You might notice that users have complained about the same issue across multiple interviews over two years—something you never would have connected otherwise.

Step 3: Ask Smart Questions

When you need information, you upload relevant files to ChatGPT and ask specific questions. The key trick Rachitsky shares: always tell ChatGPT to only use information from your uploaded files. This prevents the AI from making things up and keeps your answers grounded in your actual work.

For example: "Based only on the attached post-mortems, what are the top 3 reasons our launches have been delayed?"

Real Examples

Example 1: Preparing a Strategy Brief Instead of spending hours re-reading old documents, a PM uploads their last 12 months of meeting notes and asks ChatGPT to identify recurring themes and unresolved questions. What used to take a full day now takes 30 minutes.

Example 2: Learning From Past Launches Before a major product launch, a PM uploads previous launch post-mortems and asks: "What went wrong in past launches that we should avoid this time?" The AI surfaces specific, documented lessons rather than relying on foggy memories.

Example 3: Onboarding New Team Members A team shares their collective second brain with new hires, who can ask questions like "What has this team learned about our enterprise customers?" and get answers drawn from years of accumulated knowledge.

Product managers using this system report completing research 2-3 times faster than before, with the added benefit of surfacing connections they would have missed entirely.