Knowledge Management and Organizational Learning Analysis

Model and evidence

Knowledge is distributed across people and artifacts. Transactive memory tracks who knows what; it develops through expertise distribution, familiarity, and communication and predicts team outcomes [EV-KM-2026-3E80]. Knowledge integration mediates performance and psychological safety moderates relationships [EV-KM-2026-91BC].

Counterevidence and interpretations

Tacit knowledge may resist codification. Expertise directories can become stale. Dependence on experts creates bottlenecks, while exhaustive externalization is costly and lossy. “Knowledge sharing” is not identical to use.

Reusable methods and outcomes

Expertise-location accuracy; time to knowledgeable person; knowledge-integration quality; bus factor; psychological safety; retention after turnover; provenance and decay audits.

Coverage, gap, and challenged assumptions

KM explains organizational memory and why artifacts cannot replace relationships. It rejects the assumption that tacit knowledge can usually be fully externalized. It does not prescribe detailed software verification or interface semantics.

Discriminating experiment

Compare artifact-only handoff, expertise-routing, and combined conditions. Measure correct continuation, interruption burden, and resilience when the original expert is unavailable.

Confidence and limits

High mechanism overlap; Medium software-specific transfer confidence.

Ten future questions

1. Which knowledge is economical to externalize?
2. When is routing superior to documentation?
3. How fast do expertise directories decay?
4. Does psychological safety change clarification behavior?
5. What artifact signals reveal hidden tacit dependencies?
6. Can AI infer expertise without surveillance harm?
7. How does turnover alter handoff value?
8. Which knowledge-integration measure predicts performance?
9. When does codification destroy useful nuance?
10. Does the proposed profile outperform combined TMS and artifact practice?

Completion assessment

Bounded rapid map complete; intervention evidence in software teams remains debt.