Amey-Thakur Knowledge transfer
Knowledge concentrated in one person is a risk that materialises the week they leave. Transfer is not a document handover; it is building enough understanding in others that they can handle novel situations.
Method
- Identify the concentrations first. What only one person can do, found by asking who would be called at three in the morning (see service-catalog).
- Transfer by doing, not by telling. The receiver performs the work with the expert observing, which surfaces the tacit knowledge that explanation omits.
- Start with the failure modes. How it breaks and what to do is the knowledge most needed and least documented (see runbook-writing).
- Write the why, not only the how. Steps can be documented; understanding why the steps exist is what handles the unanticipated case (see architecture-decision-records).
- Verify by having them work alone. Transfer is complete when the receiver handles a real case unaided, not when the session ended.
- Spread across more than one person. Transferring to a single successor recreates the same risk immediately.
- Do it before it is urgent. Transfer during a notice period is rushed and incomplete (see project-closure).
Boundaries
Transfer takes real time from both people and competes with delivery. Some expertise is genuinely deep and cannot be transferred quickly. Documentation supports transfer but does not achieve it, since reading does not produce judgement.