How Do You Extract Tacit Knowledge From an Expert Who "Just Knows" the Job?

Published 11 August 2026

Philosopher Michael Polanyi summed up the whole problem in five words in 1966: we know more than we can tell.

Key Findings

Philosopher Michael Polanyi introduced the concept of tacit knowledge in his 1966 book The Tacit Dimension, built around the observation that people know more than they can explicitly articulate, a phenomenon later named "Polanyi's Paradox" by economist David Autor in 2014.
Tacit knowledge is distinct from explicit knowledge (facts and procedures that can be written down directly): it covers judgment, pattern recognition, and intuitive "know-how" that experts often can't fully verbalise on request, even when they can perform the task reliably.
Researchers Nathan and Petrosino connected tacit knowledge directly to the "expert blind spot": as experience accumulates, decision processes that once required conscious, step-by-step reasoning become compressed into fast, automated pattern recognition, meaning the expert "knows the answer at a glance" but can't easily reconstruct the reasoning behind it on request.
Cognitive task analysis is an established field of methodology specifically developed to extract this kind of expert knowledge, using structured interview and observation techniques rather than simply asking an expert to explain their process directly, since a direct explanation often only surfaces the conclusion, not the reasoning steps that led there.

Why Can't You Just Ask an Expert to Explain What They Know?

Because a large part of genuine expertise isn't stored as something the expert can consciously retrieve and describe. Polanyi's core observation, "we know more than we can tell," describes exactly this gap: an experienced person can ride a bicycle, recognise a familiar face, or diagnose a problem at a glance, but struggles to produce a complete, step-by-step account of how they did it. Asking "can you explain how you do X?" often only returns the final judgment, not the intermediate reasoning, because that reasoning has become automated and largely unconscious with enough repetition.

What's the Difference Between Explicit and Tacit Knowledge in a Training Context?

Explicit knowledge is what's already written down or easily stated: policies, defined steps, specific facts. This is the easy part of course creation, since it just needs organising, not extracting.
Tacit knowledge is judgment, pattern recognition, and intuitive know-how: knowing which of five plausible causes is actually the problem, sensing something's off before it's measurable, adjusting technique based on subtle feedback. This is the part a standard "please write down what you know" request usually misses entirely.
Most SME-authored training content over-represents explicit knowledge and under-represents tacit knowledge, simply because the explicit part is easier to ask for and easier for the expert to produce on demand.

How Do You Actually Extract Tacit Knowledge From an Expert?

Ask about specific past incidents, not general principles. "Walk me through the last time something went wrong and how you caught it" surfaces judgment in a way "what do you look out for?" doesn't.
Ask what they noticed before they could explain why it mattered. Experts often sense a problem before they can articulate the reasoning, and this sequence itself is useful training content.
Show them a real or realistic example and ask them to think aloud while evaluating it, rather than asking them to describe their process from memory in the abstract.
Ask what a less experienced colleague did differently in a similar situation, and why the expert's approach was better; contrast surfaces judgment that a description of "the right way" alone often doesn't.
Observe them doing the actual task where possible, rather than relying solely on a conversation about the task, since some tacit knowledge only becomes visible in the doing.
Ask "what would you tell someone not to do?" as often as "what should they do?" — the mistakes an expert instinctively avoids often reveal the pattern recognition behind their expertise more clearly than a description of correct practice does.

Cognitive Task Analysis: A More Structured Approach

For genuinely complex or high-stakes tacit knowledge, structured cognitive task analysis techniques go further than a standard interview. These typically involve presenting the expert with realistic scenarios and probing their decision-making at each step, rather than asking for a general description after the fact. This is more time-intensive than a standard SME interview, and is usually reserved for skills where the judgment involved carries real consequences (safety-critical decisions, complex diagnostics), rather than applied to every piece of training content.

Explicit Knowledge Extraction vs. Tacit Knowledge Extraction

Factor	Explicit knowledge	Tacit knowledge
How it's typically captured	Direct request: "write down the process" or "what's the policy?"	Indirect: incident-based questions, think-aloud observation, contrast questions
SME's ability to produce it on demand	High — usually already knows how to describe it	Low — often can perform the task but struggles to fully explain it
Risk if extraction relies only on direct questions	Low — this is where direct questions work well	High — direct questions ("explain your process") often only surface the conclusion, not the reasoning
Best extraction method	Structured interview or existing documentation	Incident recall, contrast questions, or observation

FAQ

How do you extract tacit knowledge from an expert who "just knows" the job?

Ask about specific past incidents rather than general principles, use contrast questions ("what did a less experienced colleague do differently?"), and where possible observe them performing the actual task, since tacit knowledge often surfaces indirectly rather than through a direct request to explain their process.

What is tacit knowledge, and how is it different from explicit knowledge?

Explicit knowledge is what can be directly stated: facts, defined steps, written policies. Tacit knowledge, a concept introduced by philosopher Michael Polanyi, covers judgment, pattern recognition, and intuitive know-how that experts often can't fully articulate, even though they reliably apply it.

Why does asking an SME to "explain their process" usually miss important content?

Because for genuinely experienced practitioners, much of their decision-making has become automated pattern recognition rather than conscious, step-by-step reasoning. A direct request to explain often only returns the final judgment, not the intermediate reasoning that's actually become tacit.

What's cognitive task analysis, and when is it worth using?

It's a structured methodology for extracting expert judgment, typically using realistic scenarios and probing questions during the decision-making process itself, rather than relying on a description after the fact. It's more time-intensive than a standard SME interview, so it's usually reserved for high-stakes or safety-critical skills.

Can tacit knowledge ever be fully captured in written training content?

Not entirely, according to the underlying theory. Polanyi's central claim is that a wholly explicit account of tacit knowledge is, in principle, unachievable. Extraction techniques can surface much more of it than a direct request would, but some of what makes an expert reliably good at their job likely remains genuinely difficult to fully articulate.

Where This Extraction Approach Still Has Limits

Even good extraction technique doesn't turn tacit knowledge into something a novice can apply immediately after reading about it. Some of what makes tacit knowledge tacit is that it's built through repeated exposure and practice, not just transmitted information. Training built from well-extracted tacit knowledge can accelerate a novice's development considerably, but it typically still requires the novice to build their own experience before that judgment becomes reliable for them too.

Want to see how incident-based interview content could be structured into a course? Start free.