How to capture a retiring founder's knowledge
The method in short: start at the retirement announcement, not later; map what only the founder knows; capture primarily from his daily work (meetings, decisions) rather than from writing assignments; add structured deep-dive sessions per domain; and verify the result with real team questions while he can still correct it. The order matters less than the start date.
Before you start
You need three things before step one: a founder who is willing (nothing works against him), a rough departure date (it sets the budget of time), and the acceptance that this is a program of months — not a farewell-week activity. If the departure is closer than six months, run the same steps compressed and adjust expectations honestly.
The steps
Step 1 — Map what only he knows
List the questions only the founder can answer. Source them from the team: 'what do you walk to his office for?' Twenty to forty questions is normal. Group them into domains — typically decision history, client relationships, pricing, suppliers, and the unwritten rules.
Pitfall: Asking the founder to make this list himself. He can't see his own tacit knowledge — the team's interruptions are the better map.
Step 2 — Prioritize by damage, not by interest
Rank the domains by what it costs if the knowledge is gone: frequency of use times cost of a wrong answer. Client and pricing knowledge usually outranks the war stories, however good the stories are.
Pitfall: Letting the founder's enthusiasm drive the order. The history he loves telling is rarely the knowledge the company bleeds without.
Step 3 — Capture from use, not from homework
Record and ingest the meetings where he decides and explains — that is where the reasoning is live. Writing assignments fail with founders almost universally; speech in context is the native format of this knowledge.
Pitfall: Scheduling 'documentation Fridays'. They get cancelled by week three, and what does get written is the sanitized version, not the reasoning.
Step 4 — Run deep-dive sessions per domain
Structured interviews, one domain at a time, led by someone who prepared: walk through real cases, ask why at every fork, chase the exceptions ('when would you not do this?'). Sessions of 60–90 minutes, spaced over weeks — knowledge surfaces in layers.
Pitfall: One marathon session. The second pass over a domain, a week later, reliably yields what the first pass couldn't — skipping it forfeits the depth.
Step 5 — Verify while he can still correct
Make the captured knowledge askable and let the team use it on real questions before the departure. Every wrong or incomplete answer found now is a free correction; the same gap found after departure is permanent.
Pitfall: Treating capture as done at the leaving drink. The last month of verification is worth more than the first month of capture.
How MentX compresses this
MentX is a personality-aware knowledge base: it captures the knowledge of a company's key person — from meetings and a guided intake track — into a living, temporal knowledge graph with full source attribution. MentX runs this method as a product: Connect is step 3 (meeting ingest, automatic), Deepen is steps 1, 2 and 4 (the guided intake track, led by us, personality-aware), and DigiME makes step 5 natural — the team converses with the founder's digital knowledge copy while he is still there to veto and correct. Every statement lands in a living, temporal knowledge graph with full source attribution.
Checklist
- Start date set — at announcement, not 'after the busy season'
- Question list sourced from the team (20–40 questions)
- Domains ranked by damage, not anecdote value
- Meeting capture running on his real work
- Deep-dive sessions scheduled, spaced over weeks
- Verification loop live before the farewell