The Mission

Organisations lose their most valuable knowledge when experienced people leave, retire or change roles. Much of it was never written down: it sits in people’s heads, in meetings and in informal conversations. In consulting and IT service firms, where we assume 11 to 15% of staff leave each year, this is a recurring and costly problem. Existing tools such as intranets, wikis and AI search help people find what is already documented, but they cannot reach knowledge that was never captured. KnowledgeFlow closes this gap by helping companies keep the knowledge employees would otherwise take with them when they leave or change roles. It does this through short, AI-guided interviews. The result is a reviewed article that is connected to the company’s existing knowledge in Microsoft 365, so successors, new joiners and project teams can search both in one place and see the source behind every answer. We focus on consulting and IT service firms with 50 to 499 employees in the Netherlands and Germany. Our survey of 28 professionals showed that 71% see value in capturing undocumented knowledge. Our mission is to make a handover reliable, fast and routine, so that when a person leaves, their knowledge stays.

The Challenge

When an experienced employee leaves, retires or changes role, the knowledge they built through experience leaves with them. Companies are good at documenting policies (83% do), but far less good at documenting what actually makes work function: only 32% document cultural norms and 43% client relationships (SHRM, 2025). In consulting and IT service firms, where we assume 11 to 15% of staff leave each year, a firm with 200 employees faces 22 to 30 departures a year, plus internal role changes. Replacing one employee can cost between half and twice their annual salary, mostly through lost productivity and knowledge (McFeely & Wigert, 2019). The knowledge that remains is also hard to find. Employees spend on average 2.8 hours a week searching for or asking about information (APQC, 2021). Our own survey of 28 professionals confirms both sides. 57% said knowledge that colleagues need is not being documented, 75% of those able to judge said important knowledge had at times been unavailable after someone left, and 79% had spent at least 15 minutes a day searching for internal information. The cause is rarely a lack of tools: people named unclear responsibilities (43%) and lack of time (36%) far more often than inadequate tools (29%). Documentation is simply one more task on top of a full workload, so it is done late and in a rush, or not at all. An interview with a former board member of a student association illustrated this: his handover took about 40 hours and was still incomplete. Existing knowledge-management and AI-search tools make documented information easier to find, but in the competitor reviews we analysed we found no tool addressing the capture of knowledge that was never written down.

The solution

KnowledgeFlow is an AI powered knowledge platform that captures expertise before it is lost and makes it easy to find afterwards. It works in three steps. 1. Capture. When an employee is about to leave or change role, an AI-guided interview asks them structured questions about their projects, decisions, contacts and rules of thumb. This replaces the rushed handover document with a short conversation that takes the effort off the person who is leaving. 2. Review. The AI turns the interview into a draft article. The departing expert or their team lead reviews, corrects and approves it before it enters the knowledge base, so what is stored can be trusted. 3. Retrieve. Successors, new joiners and project teams ask questions in an AI assistant that searches both the newly captured knowledge and the company’s existing documents. Every answer shows its source and a confidence label so users can check it. A dashboard shows where knowledge is at risk, for example which departing employees have not yet been interviewed. Where we are: We have built a working prototype and tested it with first users. They completed the search and capture tasks without help, and most trusted the answers. The review step is the weakest part, so editing, approval and a visible published status are our next priority. The connection to Microsoft 365 (Teams and SharePoint) and the pilot set-up are planned, not yet built. Our first customers would start with a two-month paid pilot on one real handover, set up by our team.