# Tribal Knowledge Tribal knowledge is the unwritten, undocumented information that exists only in the heads of experienced people within a team or organization. It includes the workarounds, contextual decisions, unspoken rules, historical reasoning, and "how things actually work" knowledge that never makes it into documentation. It's called "tribal" because it spreads the way knowledge spreads in a tribe: orally, through proximity, through apprenticeship. You learn it by being around long enough, asking the right people, or making the same mistakes someone else already made. ## Why it forms Tribal knowledge isn't a failure of discipline. It forms naturally because: - Some knowledge is hard to articulate. [[Tacit knowledge]] resists documentation by nature - Documentation takes time, and delivery pressure always wins in the short term - People don't realize what they know is unique to them. Their context feels "obvious" - Processes evolve faster than documentation. The real process diverges from the written one - Knowledge sharing isn't incentivized. In many cultures, being the person who knows things is a form of job security ## Why it's dangerous Tribal knowledge is a silent form of organizational risk: - **[[Bus factor]]**: when the knowledge holder leaves, retires, gets sick, or switches teams, the knowledge vanishes. The lower the bus factor, the higher the exposure to tribal knowledge loss - **[[Knowledge Decay]]**: even when the person stays, their memory of decisions and reasoning degrades over time - **Onboarding bottleneck**: new team members can't get up to speed without access to the right people, and those people become bottlenecks - **Inconsistency**: different people develop different tribal knowledge, leading to contradictory practices across teams - **Invisible dependencies**: nobody realizes the organization depends on tribal knowledge until it's gone ## Extracting tribal knowledge The goal isn't to eliminate tribal knowledge entirely (some tacit knowledge will always exist) but to reduce the organization's dependency on it: - Pair work and knowledge transfer sessions surface what people know but haven't written down - Post-mortems and retrospectives capture decision reasoning before it fades - Recording "why" alongside "what" in documentation preserves context - Regular rotation of roles and responsibilities forces knowledge to spread - AI-assisted knowledge capture ([[Agentic Knowledge Management (AKM)]]) can help by prompting people to document decisions and reasoning in real time ## The AI angle Tribal knowledge is the hardest knowledge type to feed into AI systems. AI needs [[Explicit knowledge]] to work with, and tribal knowledge is, by definition, not explicit. Organizations that don't invest in extracting and structuring tribal knowledge will find their AI systems operating with a fraction of the organizational intelligence that experienced humans have access to. This is one of the core reasons [[Enterprise Knowledge Management (EKM)]] matters. Without it, AI gets the documented 20% while the undocumented 80% stays locked in people's heads. ## References - ## Related - [[Enterprise Knowledge Management (EKM)]] - [[Knowledge Management (KM)]] - [[Tacit knowledge]] - [[Explicit knowledge]] - [[Bus factor]] - [[Knowledge Decay]] - [[Knowledge Drain]] - [[Onboarding]] - [[Agentic Knowledge Management (AKM)]] - [[Information silos]] - [[Knowledge Workers]]