Knowledge transfer for experienced people
Tacit captures that knowledge before it walks out the door: through coaching conversations with an AI, checked against your own data, and approved by the expert who owns it.
Tell me about the last time you gave a customer a delivery date that wasn't the standard one. What did you see?
That was a contractor. When their site manager calls himself, the job is usually behind schedule. So I plan an extra week, without asking.
If a contractor's site manager calls in person, add one week of slack to the delivery date.
Sixty-one. Thirty years in the company. He is the only one who knows why one customer always gets a different lead time, which supplier to call when the usual one fails, and what the price list doesn't say.
In two years, Marc retires. If he is out tomorrow, that knowledge is gone today.
That dependence is a risk on every balance sheet, and a discount in every acquisition or succession.
The insight
People rarely hide knowledge out of bad will. They hold back for reasons you can design for.
"If I explain this, I'm replaceable."
We tie the process to something real for the expert: a mentor role, a landing path, recognition.
"Nobody ever asked how I do it."
We start with their career and their proudest cases. Being asked becomes the recognition.
"It's just gut feeling."
We ask about concrete situations, not abstract processes. That is where the reasoning surfaces.
"What happens with this?"
Full transparency from session one. The expert sees and approves everything that is kept.
How it works
A human guide builds trust. The AI prepares, follows up, structures and tests between sessions.
Career reflection
Which customer did you almost lose, and save? People love telling these stories. The expertise is in them.
Case deep-dives
What did you see, what did you weigh, and what did you deliberately not do?
Meet the data
You handle this type of customer differently every time. What do you see that the system doesn't?
Correct, don't explain
The AI proposes how you reason. You fix what's wrong. Easier than starting from zero, and it feels like respect.
Validate and sign
The expert approves the result. It becomes their playbook, with their name on it.
What you get
We replay the expert's past cases with an agent and compare its decisions with theirs. Where they differ, the next interview question appears.
Decision rules and heuristics, in the expert's own words.
A catalogue of the cases where the standard process doesn't apply.
How the work really flows, ready for training or automation.
The share of past decisions an agent reproduces, measured after every round.
Open about the goal from the very first session.
Separate layers for personal reflection and company knowledge.
The expert approves everything that is kept and shared.
Something in return: a mentor role, a landing path or a bonus.
Pilot · autumn 2026
One expert, one role, a short series of sessions. You keep the playbook; we learn how to make the method better.
At Tectonic on 21 October?
Scan, get interviewed for five minutes about your own job, and receive your mini-playbook by email.