Knowledge Interview Techniques: Extracting Expertise Through Structured Conversation

A knowledge interview is a structured conversation designed to extract tacit knowledge from a subject matter expert by asking questions that bypass the expert's tendency to answer at the wrong level of abstraction. The expert defaults to answering at the conceptual level — 'we evaluate risk holistically.' The interviewer needs the operational level — 'walk me through the last three risk evaluations you did and tell me everything you noticed and decided along the way.'

The primary technique is critical incident analysis: asking the expert to narrate specific past instances rather than describe general procedures. The expert who cannot articulate a general rule for how they triage support tickets will narrate a specific ticket triage in detail, and in that narration reveal the decision framework they use without being able to state it abstractly. The narration surfaces the knowledge. The interviewer's job is to recognize it and ask the probing questions that fill in the gaps.

1. Phase 1: Scope Setting (10 minutes) Open by defining the knowledge domain to be covered and the session structure. 'Today we are focused on your process for handling enterprise escalations — specifically the judgment calls you make in the first 24 hours. I am going to ask you to walk me through specific past cases rather than describe the process generally. The goal is to capture the things that don't make it into the written procedure.' This primes the expert for the level of specificity required.
2. Phase 2: Critical Incident Elicitation (40 minutes) Ask the expert to recall 3–4 specific recent instances of the knowledge domain in action. 'Tell me about the last enterprise escalation that you handled personally — start from when you first became aware of it.' Follow with probing questions: 'What did you notice first? What did that signal to you? What did you rule out and why? When did you decide to escalate versus handle it yourself?' The expert will try to answer generically — redirect to the specific case every time.
3. Phase 3: Edge Case Mining (20 minutes) 'Tell me about a case that didn't go according to the standard process — where the normal approach didn't work or you had to improvise.' Edge cases are where the most valuable tacit knowledge lives. The standard process handles the common cases. Expert judgment handles the exceptions. Capturing the exception-handling logic is capturing the expertise that differentiates the expert from a practitioner who merely follows the documented procedure.
4. Phase 4: Principle Extraction (20 minutes) After 3–4 specific cases have been discussed, ask the expert to reflect: 'Looking at those cases we discussed, what patterns do you notice in how you approach these situations? What are the two or three things you always check first?' Some experts will be able to articulate principles at this point that they could not have stated at the beginning of the session. Others will still answer abstractly — in which case, the principles are in the narrations you have already captured, and the extraction happens in the synthesis phase after the interview.

The synthesis phase happens after the interview, not in the interview. The interviewer reviews the transcript or recording, identifies the decision patterns that appear across multiple cases, and writes them as explicit rules. 'When [condition X] is true and [condition Y] is also true, the expert consistently chooses [action Z].' These extracted rules are then validated with the expert: 'Based on our conversation, I extracted these five decision rules — do these accurately capture how you approach this?' The expert can confirm, correct, and add nuance. This validation step is critical — the interviewer's extracted rules are hypotheses, not facts, until the expert confirms them.