What does it mean that the team's knowledge graph?

A team is smarter than any individual member — but only if it knows who knows what. Transactive memory systems are the meta-knowledge infrastructure that makes collective expertise navigable.

Example: A DevOps team of eight at a healthcare SaaS company had a recurring problem: every time an unfamiliar production issue arose, the team's first thirty minutes were spent figuring out who might know something relevant. The on-call engineer would post in Slack: 'Anyone know how the auth token refresh works?' Three people would respond with partial, sometimes contradictory, answers. Meanwhile, the one person who had built the auth system two years ago was offline, and nobody remembered that she was the right person to ask. The team lead, Priya, realized the problem was not lack of knowledge — the team collectively knew everything it needed. The problem was lack of knowledge about knowledge. Priya created what she called the 'expertise map' — a living document that listed every major system component, the primary expert for each, the secondary expert, and the location of the relevant documentation. She updated it quarterly and reviewed it at each team retrospective. Within two months, mean time to resolution for production incidents dropped by forty percent. The team had not learned anything new. It had learned where its existing knowledge lived — and that meta-knowledge was the bottleneck that had been invisible for years (Wegner, 1987).

Try this: Build a transactive memory map for your team. Create a matrix with system components, processes, or knowledge domains as rows and team members as columns. For each cell, use a simple rating: E (expert — deep knowledge, can solve novel problems), K (knowledgeable — can handle routine issues), F (familiar — knows enough to ask the right questions), or blank (no significant knowledge). Fill in your own row first, then ask each team member to fill in theirs. Compare the results. Look for three patterns: (1) Single points of failure — domains where only one person is rated E. (2) Knowledge deserts — domains where no one is rated above F. (3) Perception gaps — cases where your rating of someone else's expertise differs significantly from their self-rating. Share the completed map with the team and discuss: 'Where are we most vulnerable if someone leaves?'