One resignation away

Every operating business has at least one person who is quietly holding it together. The one everyone messages when something breaks. The one who knows why the pricing exception exists, which customer cannot go on the standard contract, and how month-end actually closes once the system throws its usual error. Things work because that person carries the operation in their head.

We do not call that a documentation gap. We call it where your differentiation lives. The know-how that makes a company hard to copy, the judgment built up over years of running the work, almost never sits in a system. It sits in a few people. And the same know-how that makes you hard to copy is the know-how most likely to walk out the door.

This matters more now than it did five years ago, because most companies are buying the same AI as their competitors. The model is a commodity. The vendor is a commodity. The part that stays yours is how you actually operate, and right now a large share of it is undocumented and one resignation away from gone.

Why it looks like strength #

The reason this risk stays invisible is that it wears the costume of reliability.

The person is good, so nothing sounds an alarm. A dependable operator makes the dependency feel safe. Output is steady, escalations are rare, the org chart looks fine. Nobody prices a single point of failure that has not failed yet.

It gets defended as tribal knowledge. "That is just how Dana does it." Whenever we hear a sentence like that in an engagement interview, we write it down. It almost always marks a workflow that runs on judgment no one has ever been asked to externalize.

The documentation that exists is theater. There is a wiki. It is two years stale. The runbook covers the easy path and none of the exceptions, which are the entire reason the person is valuable. And lately there is a new version of the same wishful thinking: the assumption that an AI tool will absorb the know-how on its own. It will not. Generic tools encode the commodity. They do not encode the part that is specifically yours, because that part was never written down for them to learn from.

What it costs, sized #

Take a specialty manufacturer doing $120M in revenue. One estimator has priced the complex jobs for eleven years. His quotes carry a win rate and a margin no one else on the team can reliably reproduce, because the logic is a blend of material knowledge, customer history, and instinct he has never had to explain.

Put a number on that. If he leaves, the realistic outcomes are a six to twelve month dip in quote quality, a measurable hit to win rate on the high-margin work, and a scramble to reverse-engineer his judgment from old quotes. Against the margin on that job class, the exposure runs well into seven figures. None of it appears on a risk register. It is filed under "Dana is great," which is true, and beside the point.

The honest first move #

You do not fix this by trying to document everyone. Most of what people do is commodity work a generic tool already handles. The real target is narrow: the few workflows that are both genuinely differentiated and dependent on one person.

So the first move is a read, not a project. Find the handful of workflows where the company's edge lives in someone's head. Size what each one would cost you if that head walked. Decide which are worth formalizing now. Only then does it make sense to build, to encode the actual know-how into a system the company owns and can extend, so the judgment outlives the tenure.

That sequence matters. Capture the wrong workflows and you have an expensive wiki. Capture the right ones and you have turned your most fragile asset into a durable one.

One question for the owner #

If you want a single question that tells you whether you have a key-person problem, ask the person who runs the company:

If your most indispensable operator gave notice tomorrow, what would it cost us, and how much of what they know exists anywhere but in their head?

If you have a defensible answer, you are in a small minority. If the honest answer is "I would rather not think about it," that is the problem, and it is quietly compounding.

What we do at Joust #

The Joust Focused Read on key-person risk is a short, senior-led engagement that finds the workflows where your edge lives in a few people's heads, sizes what is at risk, and tells you what is worth formalizing. Where it is worth building, we encode that know-how into a system you own, so the differentiation stops being a person and starts being an asset.

If a resignation would hurt more than you would like to admit, book a 30-minute conversation. No deck, no sales pitch. Just the conversation.

Or email Ron Davis directly at ron@joustagency.com.