The knowledge walking out of the door

Somewhere in your company there is a person who can glance at an order, or a fault report, or a customer email, and know that something is wrong before they can say what. Ask them how they knew and they will shrug and say experience. That answer is completely true and completely useless. True, because it took twenty years to build. Useless, because it transfers nothing.

Established companies run on this kind of knowledge to a degree that is easy to underestimate, because by definition it never appears anywhere you can see it. The official process is documented, more or less. The real process lives in people. Which customer to read literally and which one loosely. What the supplier actually means by that delivery date. Which alarm matters and which one cries wolf. When the standard procedure should be quietly ignored, and when it absolutely should not. Most of it sits with the longest serving people, which is to say, with the people closest to the door.

That makes this a deadline, not a condition. A condition you can live with indefinitely. A deadline arrives whether you prepared or not, and retirements do not reschedule because the timing is inconvenient. The usual response, two weeks of handover at the end, captures almost nothing, and it is worth understanding precisely why.

Why handovers fail

The standard handover asks the expert to write down what they do, or to walk a successor through it. What comes out is the official process, again. Not because the expert is holding anything back, but because expertise does not feel like knowledge from the inside. Twenty years of pattern matching compresses into intuition, and intuition presents itself to its owner as just looking. The knowledge that matters most is precisely the part that no longer feels like knowledge at all.

You can watch this failure happen in a single conversation. Ask an experienced dispatcher how they plan a difficult day, and they will describe the planning system, which you already have. Show them one specific difficult day from last winter and they will talk for forty minutes, and every minute of it will be things that are written down nowhere. That asymmetry is not a curiosity. It is the whole method, and we will come back to it.

The binder problem

The traditional answer is documentation, and most companies have tried it. Somewhere there is a wiki, a shared drive, or a shelf of binders from a quality initiative years ago. It did not work, for two reasons.

First, it captures rules, and rules are the wrong container. The expert's knowledge is not a set of rules. It is a set of cases with judgment attached, and forcing it into rule form strips out exactly the conditionality that made it valuable. Second, even good documentation fails at retrieval. It sits where nobody looks, ageing quietly, and the person who needs it at half past four on a Friday does not go reading wiki pages. Knowledge that is not present at the moment of need might as well not exist.

What changed

The reason to reopen this old problem now is that its economics have flipped. Until recently, captured knowledge was an archive. You paid to create it, and then it sat there hoping to be read. That is a bad trade, and companies were right to be unenthusiastic about it.

What changed is that captured knowledge can now be put to work. It can sit behind systems that answer with it, check new work against it, and apply it at the exact moment it is needed, inside the tools where the work already happens. The judgment of your best person, which used to reach exactly as far as their own desk, can be present in every transaction of the kind they know best. The archive became fuel. That changes what captured knowledge is worth, and therefore what it is worth spending to capture it.

Follow the exceptions

Do not ask people what they know. They cannot tell you, as every failed handover shows. Instead, watch for the moments when the knowledge surfaces on its own, because it does, in one specific and predictable place: exceptions. Hidden knowledge shows itself whenever the standard path fails. Every time someone in your company says some version of give it to Karin, she will know what it means, a flag has gone up. Collect those moments for a month and you have a map of exactly where the undocumented knowledge sits, who holds it, and what it is about. This costs nothing. It is a noticing exercise.

Then capture by cases, not by rules. Sit down with the expert and specific history. Tell me about one that went wrong. Tell me about the strangest one this year. What did you see first? An expert who cannot state a single rule can narrate cases for hours, and the cases contain the rules, implicitly, along with all the conditions the rules would have lost. Record the narration. Extracting structure from hours of recorded human talk used to be the expensive part. It is now nearly free. What remains expensive is what was always expensive: the expert's time, and the expert's trust.

The dignity question

Which brings up the part that is usually handled worst. From the expert's side, a knowledge capture project can look like a plan to make them replaceable, and when it is framed internally as reducing key person risk, that suspicion is roughly accurate. People are not stupid about this. An expert who feels extracted will give you the official process and a pleasant smile.

The honest framing is the opposite one, and it has the advantage of being true: this is how their twenty years outlasts their employment. Nothing else on offer does that. The pension does not. The farewell cake does not. A working system that carries their judgment, corrects what they would correct, and catches what they would catch is the only form in which their experience survives them. Most experts, offered that honestly, discover they want it. Done right, capture is not a threat to the expert. It is the only monument they are ever going to get.

An expiring asset

A company's real age is measured in this material. It is the one asset a young competitor cannot buy, copy, or hire quickly enough to matter. Every year of operation adds to it. But it has an unusual property for an asset: it only becomes durable if you capture it while its holders are still in the building. Experience appreciates. Access to it expires. The companies that treat those two sentences as one sentence are the ones that will still be explaining things to their competitors in ten years.