PM Knowledge Engine

Problem

Product management knowledge is scattered across podcasts, blogs, books, and tribal knowledge. PMs spend hours consuming content but struggle to systematically apply frameworks when they need them.

Solution

Built a knowledge extraction system that processes PM content (podcast transcripts, articles, books) and organizes it into a structured, searchable framework database.

How It Works

1. Ingest — Process podcast transcripts and articles through an LLM pipeline
2. Extract — Identify frameworks, mental models, and actionable patterns
3. Classify — Categorize by topic (strategy, prioritization, growth, leadership)
4. Structure — Create consistent framework cards with: name, source, when to use, steps, examples
5. Serve — Searchable interface with context-aware recommendations

Key Frameworks Extracted

• Prioritization: LNO, ICE, RICE, Opportunity Cost models
• Strategy: DHM Model, JAM Model, Horizons Framework
• Product-Market Fit: Sean Ellis Test, PMF Engine
• Growth: Product-Led Sales, Network Effects, Viral Loops

Learnings

1. LLMs are great at extraction — with good prompts, they can identify frameworks in unstructured content with high accuracy
2. Deduplication is the hard part — the same framework gets described differently by different sources
3. Context matters more than content — knowing when to use a framework is more valuable than knowing what it is