Succession Planning When One Person Knows Everything

Succession planning gets discussed as an ownership question: who buys the shares, what the business is worth, how the transition is financed. Those matter and they are rarely the thing that breaks. What breaks is that one person knows how the pricing actually works, which suppliers will bend on a deadline, and why the machine in the back is never run above a certain setting.

Start with the month-long absence

The useful exercise takes an afternoon. For each key person, list what would simply stop if they were unavailable for a month with no handover. Not reduced. Stopped.

Write it per person rather than per process, because the concentration is the finding. Most owners come out of this with two or three items they genuinely cannot delegate and a considerably longer list of things they have simply never written down, and the second list is where the cheap wins are.

Do it for yourself first if you are the owner. The answer is usually longer and more uncomfortable than it is for anyone else on the payroll, which is itself the result worth having.

Four kinds of knowledge, and what transfers

The first two categories are where capture technology has genuinely changed the economics, and together they are more of the total than people expect. The second two have not changed at all, and treating a document as a substitute for an introduction is how a handover looks complete and fails in month two.

Record, do not write

Asking a long-serving employee to document their job produces a two-page summary of what they think you want to hear, delivered three weeks late. Asking them to narrate a task while doing it produces forty minutes of detail they would never have thought to write.

Transcribe that, have it structured into a procedure, and hand it back for correction. Correcting a draft takes a fraction of the time of producing one, and the corrections are where the real content arrives, because people are far better at noticing a wrong step than at recalling a complete one.

Accuracy on names, part numbers and settings needs a check, since transcription fails on exactly that vocabulary, which is covered in where automatic transcription falls short. The format to aim for is in standard operating procedures with AI.

The archive already holds a lot of it

Before running capture sessions, look at what is already recorded. Years of email, job notes, quotes and support threads contain a great deal of how this business works, and the obstacle has always been that it was unsearchable rather than unwritten.

Making that archive answerable is now a modest project rather than a large one, and it reaches the exceptions that nobody would think to document. The practical mechanics are in building a company brain, and where the material is on paper it starts with scanning.

One caution. An archive that answers questions confidently from outdated records is worse than no archive, so date the sources and prefer recent material where they conflict. Contradictory records are a cleaning problem before they are a retrieval problem, as in data cleansing.

Judgment transfers by example

The hardest category is the decisions made by someone who has seen the situation forty times. A procedure can say what to do; it rarely captures when not to follow the procedure.

The method that works is unglamorous. Collect worked examples of real decisions with the reasoning attached, especially the ones that went against the obvious answer, and review them with whoever is learning. A dozen of those teaches more than any document, and recording the conversation means the twelve examples exist for the next person too.

This is also where handing judgment to an assistant goes wrong, because it will produce a confident answer for a situation it has no basis to judge. Keeping the person in the decision while the tool handles retrieval is the arrangement that holds, and the risk of letting it erode is in keeping your team sharp.

Frame it correctly or it will not happen

Asking someone to document everything they know lands as a request to make themselves replaceable, and the cooperation you get will reflect that. The honest framing is that the business should not depend on any single person being reachable, which is true, applies to the owner most of all, and is usually accepted when the owner goes first.

Capture as a continuous habit beats capture as an exit process. Thirty minutes a month on one task produces a complete picture within a year, and it avoids the version where a resignation letter starts the clock. The wider framing of working alongside these tools is in training your workforce to collaborate with AI.

Frequently Asked Questions