How do you retain knowledge when experienced staff leave?

What actually walks out of the door?

When an experienced person hands in their notice, the firm usually worries about their clients and their billing. The larger loss is harder to see. It is everything they know about how the work is really done, which was never written down because it never needed to be.

Think of a senior associate or manager with years at the firm. They know which clients want the numbers before the narrative, and which want a phone call before any bad news is put in writing. They know the house template has a clause that always needs changing for a particular kind of engagement. They know which regulator queries tend to follow a certain type of submission, so they add a paragraph that answers the question before it is asked. They know which colleague in another team actually understands the pension scheme rules.

None of that is in the file. It is in them. For the length of their notice period, it is still available. After that, the firm finds out what it lost one mistake at a time.

Why do handover documents fail?

The standard response to a resignation is a handover document. It is well intended and it nearly always disappoints, for reasons that are structural rather than personal.

It is written at the worst possible moment. The leaver is finishing their own work, handing over clients and thinking about the next job. Writing a thorough account of everything they know competes with all of that and usually loses.

It captures what the leaver thinks to write down, which is the explicit part: open matters, deadlines, contacts and passwords. The valuable part is tacit. Experts do not experience their judgement as knowledge. They experience it as common sense, so it does not occur to them to record it.

It is read once and then filed. A handover note is a document about the work, sitting beside the work. When the person who inherited it faces a decision months later, they will not think to look.

And it assumes a successor. Often the work is spread across several people, none of whom receives the whole picture.

Where should knowledge live instead?

The only knowledge a firm reliably keeps is knowledge that is used every day. If a check is part of how a document is produced, it survives the departure of the person who invented it. If it lives only in their habits, it does not.

This reframes the problem. Retaining knowledge is less about documentation and more about design. The question to ask is not "have we written down what this person knows?" but "does the way we do this job carry what this person knows, whoever is doing it?"

In practice that means a few specific things:

• Templates that carry the judgement. If a clause always needs changing for a certain type of client, the template should know that, not the person filling it in. Why templates rot explains how that knowledge usually decays instead.
• Rules made explicit. Client preferences, sector quirks and house style written as rules the process applies, rather than remembered by whoever happens to handle the account.
• Checks built into the flow. The extra question a senior person always asks becomes a step the draft passes through, with a named reviewer.
• Records that explain themselves. Notes and reports produced in a consistent structure, with the source of each figure kept, so a newcomer can follow the reasoning.

Doesn't searching our own documents solve this?

Partly. Tools that search across a firm's past files, precedents and correspondence are genuinely useful. A newcomer can find how a similar question was answered before, or what a client was told on a previous engagement. Knowledge retrieval covers what that looks like done properly.

But search can only return what was recorded. It will find the advice letter. It will not find the reason the senior person rewrote the second paragraph three times, or the phone call that preceded it. Retrieval is a good way to reuse written knowledge. It does nothing for knowledge that was never written.

Why does rebuilding a process capture knowledge as a side effect?

When a process is rebuilt properly, the first step is sitting with the people who do the job and mapping every input, template, handoff and check. That exercise draws out exactly the tacit knowledge a handover note misses, because it asks about the work while it is happening rather than about the person after they have decided to leave.

Questions like "why do you always change this?" and "what would go wrong if you skipped that?" surface the reasoning experts no longer notice. Once the answers are known, they are built into the templates, rules and checks the process uses. The knowledge stops belonging to one person and starts belonging to the job.

That is not the main reason to rebuild a process. The main reason is usually the hours it takes. But the effect on resilience is real, and for many firms it is the benefit that matters most a year later, when someone important moves on and the work carries on without a dip.

It also changes things for the people who stay. Juniors learn faster when the process shows them what good looks like, rather than relying on who they happen to sit near. What changes for junior staff looks at that side.

Where should a firm begin?

Pick one job that depends heavily on a small number of experienced people and imagine each of them leaving next month. Ask what would break first. That is where your knowledge risk is concentrated. How to document a process gives a method for capturing it, and standard operating procedures covers turning the result into something the team actually follows. If the same job is also eating hours every week, the audit will show you whether it is the right place to start.