Organizations Urged to Adopt 4-Plane Architecture for Tacit Knowledge Capture

2 articles · Updated · O'Reilly Media · Sep 16

A 4-plane “tacit-aware” architecture is being proposed to help organizations capture the judgment, exceptions and preventive know-how that standard documentation and retrieval systems routinely miss.
The push targets a core enterprise AI weakness: better search can surface recorded material, but it cannot recover expertise never captured—especially rare decisions, perceptual skill and team habits that leave little or no data trail.
A structured elicitation protocol centers on 4 follow-up areas—thresholds, exceptions, evidence and escalation—to turn vague expert statements into usable rules, while limiting model use to suggesting sharper interview questions.
The broader design spans capture, representation, serving and transmission, pairing documented knowledge with provenance, referrals to experienced staff and apprenticeship when expertise cannot be reduced to prose.
Evaluation should go beyond retrieval accuracy to incident replays, bus-factor audits and abstention tests, with success measured by faster proficiency, fewer repeat incidents and less dependence on 1 expert.