Building a shared knowledge lens for distributed product squads

As we scaled from 3 to 30 product squads, knowledge became fragmented. Decisions made in one squad weren't visible to others. Patterns that worked weren't captured. Lessons learned were lost.

We built a knowledge graph—not a wiki, not a document store, but a structured system that connects insights, decisions, patterns, and outcomes across all squads.

Every playbook, decision, experiment, and outcome gets tagged with structured metadata: problem space, solution approach, outcome metrics, related work. The graph automatically surfaces connections.

When a squad starts a new project, they query the graph: 'What have we learned about onboarding flows?' The system surfaces relevant playbooks, past experiments, and related decisions from across the org.

The magic is in the connections. A playbook about 'reducing signup friction' connects to experiments about 'email verification flows' and decisions about 'password requirements.' Squads see the full context, not just isolated documents.

We built this on top of Notion's API, using custom schemas and automated tagging. Every week, squads update their playbooks with new learnings. The graph grows smarter over time.

The impact? 60% reduction in duplicate work, 40% faster onboarding for new squads, and a 3x increase in cross-squad knowledge sharing. Knowledge isn't just stored—it's actively used to make better decisions.