How do I stop losing company knowledge when staff leave?
Losing knowledge is the same problem as guessing your next move. The decision logic experienced people use lives in their heads, not in the system.
Short answer: Most company knowledge leaves with staff because it is tacit: the judgment, problem fixes and decisions that live in people's heads and phones, not in documents. To keep it, capture the logic behind decisions, not just the tasks. Store it where it can be searched, then hand it to new staff and AI in a structured format.
Why your SOPs did not solve it
SOPs record what to do. They rarely record how an experienced person decides. When an engineer, agent or manager resigns, the manual stays and the judgment goes. The next hire starts from zero, and training restarts again.
A typical pattern: job records, photos of problems and the decisions made on site sit on individual phones. Nobody can search them. Nobody has time to organise them, because the team is busy with the next job.
What actually gets lost
- Why a problem was solved this way and not another way
- Which signs mean a job is about to go wrong
- Which customers, suppliers or machines need special handling
- The exceptions that never made it into a procedure
How to fix it
1. Pick one role that keeps turning over. Do not start with the whole company.
2. Collect what already exists. Photos, chat messages, job notes, emails.
3. Ask the experienced person three questions: When do you act before anyone asks? Which combinations of signals worry you? What decision will you face again next year?
4. Write each answer as Condition, Action, Reasoning.
5. Store it in one searchable place, so a new hire or an AI assistant can find it.
6. Review monthly. The first version will have gaps. Closing them is the real work.
Where AI helps, and where it does not
AI can search, summarise and organise what you have captured. It cannot invent judgment that was never written down. Capture first, then automate.
Where to start this week
Choose the person you would least want to lose. Spend 60 minutes recording how they decide one repeating problem. That is your first entry.
Related: How do I make years of old company data searchable with AI? | What is a trigger library?
Lobang Lab is a community founded by Chris Phang that helps experienced business owners turn what their people know into searchable, transferable systems with AI, so the business keeps its knowledge when staff leave or the founder steps back.