When the Expert Leaves: How to Keep Their Knowledge in the Business
Imagine an office where everyone turns to the same person. They know where the missing keys are, why the monthly report never balances first time, and which process has three extra steps nobody has written down.
Then they hand in their resignation.
The first reaction is usually practical. Who can replace them? How quickly can someone new start? What work needs to be covered this week?
Those questions matter, but they only solve part of the problem. Replacing a person is one task. Making sure their knowledge stays with the business is another.
If the leaving employee is the only person who understands how things really work, the risk is bigger than an empty seat. It affects continuity, customer service, staff confidence, compliance, and the time it takes for the next person to become useful.

The real problem is the knowledge gap
When someone leaves, a business usually sees the obvious work first.
There are tasks to reassign, clients to notify, files to locate, passwords to reset, and deadlines to protect. That creates pressure, so the handover often becomes a list of unfinished jobs.
A list helps, but it does not explain judgement.
It will not show why a supplier needs chasing before the end of each month. It may not explain which figures in the report are always wrong on the first export. It rarely captures the minor workarounds that keep a process moving.
That is where the knowledge gap appears.
A new person can be capable, experienced, and keen to learn, yet still struggle if the role depends on context nobody captured. They may know what to do next, but not how to spot when something is off. They may know where a system is, but not why the team uses it in a certain way.
The goal is not to make every person interchangeable. Good people bring skill, judgement, and relationships that cannot be copied into a document.
The goal is simpler and healthier.
This does not make someone less valuable. It makes sure they are not the only person who can take a holiday.
Ask what lives in someone’s head rather than in your systems
Every workplace has two versions of its processes.
One version is written down. It lives in software, manuals, shared folders, checklists, templates, and calendars.
The other version lives in people’s heads. It includes shortcuts, exceptions, history, warnings, preferences, fixes, and “just ask Sam” knowledge.
The second version is often more accurate than the first.
A good starting point is to ask what people know that the business cannot currently find. Look for knowledge in these areas:
recurring tasks with unwritten steps
exceptions to normal processes
passwords, access rules, or approval paths
supplier, customer, or stakeholder preferences
regular reports that need manual checking
old decisions that still affect current work
seasonal tasks that only happen a few times a year
problems that one person always solves
Some of this knowledge feels too small to document. That is usually a sign it matters.
A tiny detail can hold up a whole process. For example, a monthly report may need a manual adjustment because one category pulls through incorrectly. If only one person knows that, the next person may trust the first result and send out the wrong figures.
The question is not whether every detail deserves a long procedure. Many do not. The question is whether someone else could recover the detail when it matters.
A short note can be enough:
where the report comes from
what usually looks wrong
how to check it
who to ask if the figures still do not make sense
That small record turns private knowledge into business knowledge.
Test whether someone else could find the answer
A useful knowledge system passes a simple test: could another team member find the information without calling the person who knows it?
This test is more revealing than asking whether something is documented. Many businesses have documentation that technically exists but cannot be found, trusted, or understood.
A file buried six folders deep is not useful. A checklist last updated three years ago may be worse than nothing. A process titled `final_FINAL_v3` does not inspire confidence.
Try a practical test. Pick a common task and ask someone who does not normally do it to find the instructions. Do not let them ask the usual expert straight away. Watch what happens.
Can they find the right folder?
Can they tell which version is current?
Do the instructions match the system they are using?
Do they understand the outcome, not just the clicks?
Can they tell what to do if the normal path fails?
If they get stuck, the problem may not be the person. The system may be too dependent on memory.
This does not call for a huge documentation project. Start with the tasks where delays would hurt most. A simple page, checklist, or screen recording may solve the problem.
The best knowledge records are easy to find, easy to update, and written in plain language. They do not need to be perfect. They need to be useful during a busy day.
Identify tasks only one person knows how to complete
Single-person knowledge is common in small and growing businesses. It often develops for good reasons.
Someone becomes good at a task, so they keep doing it. They fix a recurring problem, so everyone sends it to them. They have been around the longest, so they remember why things work the way they do.
That can feel efficient until the person is sick, on leave, promoted, overloaded, or leaving.
To find the risk, map tasks against people. Keep it simple.
Question | What to look for |
Who does this task now? | The named person who usually completes it |
Who else can do it confidently? | A real backup, not someone who has watched once |
Where are the instructions? | A current, findable record |
What happens if it is delayed? | Customer, financial, legal, or operational impact |
How often does it happen? | Daily, weekly, monthly, seasonal, or rare |
This kind of map often shows hidden pressure points. A low-profile task may turn out to be business-critical. A monthly job may be fine most of the year, then cause panic when the expert is away during a deadline.
Pay special attention to tasks that combine system knowledge with judgement. These include payroll checks, billing corrections, stock reconciliations, tender submissions, compliance reporting, account access, and customer complaints.
The question is not “could someone work it out eventually?” In many cases, they could. The real question is how much time, stress, risk, and rework it would take.

Share knowledge during normal working weeks
Many businesses only talk seriously about knowledge transfer after someone resigns.
By then, time is short. The leaving employee has their own work to finish. The manager wants a handover document. The new person may not have started yet. Everyone asks for “everything you know”, which is impossible to provide in a rush.
A better approach is to make knowledge sharing part of ordinary work.
That can look like:
pairing people on recurring tasks
rotating responsibility for monthly processes
asking the expert to narrate a task while someone else records the steps
using checklists for work that often gets interrupted
reviewing key procedures once a quarter
adding “what changed?” notes after a process is updated
setting aside time after major problems to record the fix
The timing matters. Knowledge shared under pressure is often incomplete. People explain the steps they remember, not the context they use.
During a normal week, there is more room to ask better questions.
Why do we do it this way?
What usually goes wrong?
What would make you stop and check?
Who needs to know when this is finished?
What would a new person misunderstand?
These questions capture judgement, not just sequence.
They also reduce resentment. If the expert only gets asked to explain their job after resigning, the business has already left it late. If knowledge sharing is normal, it feels less like an extraction and more like part of good work.
Build handovers that explain the role, not just the open tasks
A handover should help someone understand the role. It should not only list what remains unfinished.
A task list answers “what is happening right now?” A useful handover also answers “how does this role work?”
That means including context such as:
the purpose of the role
regular deadlines and rhythms
key contacts and what they need
systems used and where records live
common problems and fixes
decisions the person can make alone
decisions that need approval
risks to watch
work that is easy to miss
current issues, not just current tasks
The best handovers are written for a capable person who does not have the same history. They do not explain every basic concept, but they do explain local knowledge.
For example, “send monthly report to finance” is a task. It may be technically correct, yet still too thin.
A better handover might say:
“The monthly report is pulled from the finance system on the first working day after month end. The export often duplicates cancelled orders, so check the cancellation tab before sending. Finance prefers the report before 12 pm because they use it in their afternoon reconciliation. If the numbers are more than 5 per cent above or below the previous month, check with operations before sending.”
That kind of note gives the next person a fighting chance. It shows timing, quality checks, stakeholder preference, and a warning sign.
It also protects the team from confusing activity with understanding.
Keep the expert involved without making them the only safety net
There is a careful balance to strike with experienced staff.
If someone holds a lot of knowledge, they are valuable. They may also feel protective of their work, especially if they built the process or rescued it many times. A clumsy knowledge transfer effort can sound like the business is trying to reduce their worth.
That is the wrong message.
Better knowledge sharing should recognise expertise. Ask the person to help improve the system, not just empty their head into a folder.
A respectful approach might sound like:
“We rely on you for this because you understand it better than anyone. We need to make sure the process is easier for others to follow, and we would like your help capturing the parts that are not obvious.”
That frames the work as leadership, not replacement.
It also helps prevent burnout. Being the only person who can solve a problem may feel rewarding for a while, but it can become exhausting. People cannot switch off if every holiday comes with calls, messages, and urgent questions.
Knowledge sharing gives experts more room to do higher-value work. It also lets them take leave without worrying that everything will stop.
Start with the knowledge that would hurt most to lose
Trying to document everything at once usually fails. It becomes too big, too vague, and too easy to postpone.
Start with the knowledge that would cause the most disruption if it disappeared next week.
Choose three to five areas and work from there. Good candidates include:
tasks tied to payment, payroll, billing, or cash flow
customer-facing processes with deadlines
compliance or reporting obligations
system administration and access
operational tasks with no clear backup
seasonal work that is easy to forget between cycles
For each area, capture the basics:
What is the task meant to achieve?
When does it happen?
Who does it now?
Who is the backup?
Where are the instructions?
What commonly goes wrong?
How do you know it has been done properly?
Who needs to be told?
This is often enough to reduce the biggest risk.
The record can improve over time. Each time someone uses the instructions, they can fix unclear parts. Each time a process changes, they can update the note. The aim is living documentation, not a perfect manual that no one touches.
Make knowledge easy to maintain
Knowledge retention fails when it becomes separate from the work.
If people have to stop, open a complicated template, write a long update, and get approval for every change, they will avoid it. The system needs to fit the pace of the workplace.
A few simple rules help:
Keep records close to where the work happens.
Use clear names and dates.
Make ownership obvious.
Remove old versions or mark them clearly.
Prefer short, current notes over long, stale manuals.
Review critical processes on a regular schedule.
Treat updates as part of finishing the task.
For example, if a monthly checklist changes, update it while the task is fresh. If a customer changes their submission preference, record it at the time. If a workaround becomes permanent, add it to the process instead of leaving it in someone’s memory.
Small updates prevent large handover problems.
They also make onboarding easier. A new starter can learn faster when the business has clear records. They still need support, but they do not have to rely on folklore to understand how things work.
A resignation should not be the first knowledge audit
When an expert resigns, the business learns quickly where the gaps are. The trouble is that the lesson arrives late.
A healthier business looks for those gaps before the resignation letter appears.
The work starts with a few honest questions.
What lives in someone’s head rather than in your systems?
Could another team member find the information without calling them?
Which tasks does only one person know how to complete?
Is knowledge shared during normal working weeks, or only during a rushed handover?
Would your handover help someone understand the role, not just list unfinished tasks?
The answers may be uncomfortable, but they are useful. They show where the business depends too heavily on one person, and where a small amount of documentation, cross-training, or process review would make a real difference.
When the expert leaves, the aim is not to pretend they were easy to replace. They probably were not.
The aim is to make sure their knowledge does not walk out with them.




Comments