The resignation letter is on your desk. Maybe it came out of nowhere, maybe you saw it coming. And while you are still figuring out how to fill the position, something else dawns on you: this person knows things nobody else knows.
How to invoice the special customer who has had custom terms since 2019. Why the machine in Building 2 needs a different startup procedure on Mondays. Where the login credentials for the supplier portal are kept. Who to call at customs when a shipment gets stuck.
None of it is written down anywhere. And the clock is ticking.
The notice period: your last chance - but a short one
Depending on tenure and contract, you typically have somewhere between four weeks and several months before the person walks out the door. That sounds like plenty of time. It is not.
Do the math: remaining vacation days have to be used up. Day-to-day business keeps running - the departing employee does not suddenly work on nothing but their handover. Sick days often come on top, sometimes garden leave. Out of an eight-week notice period, you realistically get two to three weeks in which knowledge can actually be handed over.
And one more thing: motivation drops. Someone who has mentally already moved on to the new employer no longer writes lovingly detailed manuals. That is human, not malicious. But you have to plan for it.
What is realistically achievable in that time - and what is not
A sober look pays off here, because false expectations burn through those precious weeks.
What works
- Documenting the most important processes - not all of them, just the ones nobody else can do
- Handing over access and tools - passwords, portals, licenses, admin rights
- Passing on contacts - who is the right person to talk to at customer X, at supplier Y?
- Handing over work in progress - open projects, half-finished quotes, verbal agreements with customers
- Training a successor, if one is already on board - side by side, not by document
What does not work
Most tacit knowledge is lost. Period. Fifteen years of experience - which customer gets nervous if you do not reply within a day; which part is worth double-checking; which Excel sheet is the current one and which is only maintained out of habit - none of that ends up in a handover document. Not because the person is holding back, but because they do not even know that they know it.
A handover document answers the questions the writer considers important. The questions the successor will have three months from now are different ones.
So: save what can be saved - but prioritize brutally.
The week-by-week handover plan
Let us assume you have four effective weeks. Here is how to divide them:
Week 1: A map, not the details
Sit down with the person - an hour is enough to start - and make a list: Which tasks do they handle regularly? Which ones rarely but critically (year-end closing prep, maintenance intervals, audits)? Which tasks would immediately catch fire without them?
Sort that list by a single criterion: Can anyone else in the company do this? Everything a colleague also masters gets cut. What remains is the real risk list - usually five to ten items, not fifty.
In parallel: sort out access. Which systems, which passwords, which admin rights are tied to this person? It is tedious, but quickly done - and it is the item that gets most expensive after the last working day if it is missing.
Week 2: The critical processes - in conversation, not on paper
Take the risk list and work through the items one by one - ideally like this: the person demonstrates the process once, a colleague or successor watches and asks questions, someone captures the result. Screenshots of the key screens, notes on the exceptions ("if the customer is from Austria, then...").
Why demonstrate instead of write? Because demonstrating surfaces the things that run on autopilot in someone's head. "Oh right, first you have to quickly check the other list" - that sentence never comes up while writing, always while showing.
Week 3: Work in progress and relationships
Now the open items: Which projects are running? What has been promised to customers verbally? Where are there agreements that are not written down anywhere - the discount the boss once promised on the phone, the late payment terms quietly tolerated for a long-standing customer?
And the contacts: for key customers and suppliers, a joint handover email or a short three-way phone call is worth it. That costs twenty minutes per contact and saves the successor months of building trust.
Week 4: Test run and closing the gaps
The hardest and best exercise: the departing employee is "unavailable" for one day - the successor or the team works with what has been documented. Every question that comes up is a gap. Use the final days to close exactly those gaps.
It feels uncomfortable. It is still the only honest test of whether the handover worked.
Why structured interviews save more than "just write everything down"
The reflex in many companies is: "Please document everything before you leave." The result is almost always the same - a twelve-page Word document that describes the obvious and leaves out what matters.
That is not for lack of goodwill. It is how knowledge works in the head: what you do every day, you consider self-evident. Nobody writes down that "of course you only release the order after quickly checking with Hanne" - that goes without saying. For the successor, it does not.
Good questions bring out what writing down never surfaces:
- "Walk me through a perfectly normal Tuesday."
- "What was the last thing that went wrong - and how did you notice it was going wrong?"
- "If you were on vacation for three weeks: what would pile up? Who would call you, and about what?"
- "At which step do you deviate from the official procedure - and why?"
A conversation built on questions like these unearths more in one hour than three days of solitary documenting. The person being interviewed does not have to know what is important - the questions find it.
This is exactly the principle that software solutions now build on as well: at wissa.ai, an AI conducts these structured interviews via chat, 15 to 20 minutes per person, and turns them into searchable process documentation - including a heads-up when two people describe the same process differently. If the interviewee mentions "Max actually handles that part," Max is automatically invited to the next interview. For the acute resignation case, this is faster than any manual documentation - but the real leverage lies elsewhere.
The better answer: secure knowledge before anyone resigns
Because let us be honest: everything in this article is damage control. The notice period is the worst possible time to secure knowledge - time is short, motivation is at rock bottom, and what has grown over fifteen years does not fit into four weeks.
The real mistake happened long before the resignation: the knowledge was allowed to live in a single head for years. A shout across the hallway instead of a documented process. The manual from 2021 that nobody opens anymore because everything has changed since.
The sustainable solution is unspectacular: collect knowledge regularly while everyone is still around - and in a way that happens on the side, not in workshop days nobody ever has time for. If you already know today which processes depend on which person (keyword: bus factor), the next resignation will not catch you unprepared. Then the notice period becomes what it should be: an orderly handover - not a rescue mission.
If you want to tackle this systematically: wissa.ai continuously interviews your employees via AI-led interviews and keeps your process documentation up to date on its own - every statement with a source, every change traceable. The beta waitlist is open: Join the waitlist now - €149/month in the early beta, free for the people being interviewed.
FAQ
Can I require an employee to document their knowledge before leaving?
A proper handover is generally part of an employee's duties up to the last working day - when in doubt, check the details with your employment lawyer. More important than the legal question, though, is the human one: someone who leaves on good terms and experiences the handover period as respectful will hand over far more than the bare minimum. Pressure produces twelve-page documents with no substance.
What about knowledge when the employee is put on garden leave?
Garden leave makes things tight, but not impossible. Find out whether a cooperative handover is possible in the first days of the leave - many employees are willing if the separation is fair. Priority then goes to access credentials, work in progress, and a single structured handover conversation covering the most critical processes.
Is two weeks of overlap between successor and predecessor enough?
Two weeks side by side are worth more than any document - but only if the right situations come up during that time. The month-end close, the escalation with the major account, the special case in shipping: whatever does not happen in those two weeks does not get handed over either. So always supplement the overlap with structured conversations about the rare, critical cases.
How do I identify which knowledge is truly critical?
Ask one simple question per task: could someone else in the company take this over tomorrow - without calling the person? If not, the task goes on the risk list. Usually five to ten truly critical items remain per person. That is where you focus the scarce time.
What if the notice period is already half over?
Then compress the weekly plan: secure access (one day), create the risk list (one hour), and for each critical process hold a single session in which the person demonstrates the workflow while someone takes notes and screenshots. The ideal of completeness is out of reach at that point - but you will still close the most expensive gaps.