In short
- Knowledge has feet: it leaves with employees unless you anchor it structurally instead of keeping it in people's heads and scattered documents.
- AI is not a solution for messy knowledge, but a magnifying glass. What is wrong becomes more visible and more expensive.
- Companies that take ownership and governance seriously accelerate onboarding, multiply expertise and survive staff turnover without losing knowledge.
During one of my first job interviews, more than twenty years ago now, an entrepreneur said something that has stayed with me. His company did not sell products, he said, but knowledge. And knowledge is fundamentally different from, say, a machine that makes plastic bottles.
"That machine," he said, "doesn't have feet. But you do."
His point was clear: when you invest in knowledge, you invest in people. And people can leave. They can take their knowledge with them to a competitor and make it pay off there all over again. That is why happy employees are not a nice-to-have in a knowledge business, but a strategic necessity.
Twenty years later, that statement is still just as relevant. Perhaps even more so, because a new player has entered the picture, one that does not solve the problem, but does expose it: AI.
Knowledge that only exists in people's heads leaves as soon as those people do. That sounds obvious, but we rarely act accordingly. In a knowledge business, we invest heavily in people (training, coaching, learning on the job, ...) and yet we are surprised time and again when knowledge disappears when someone leaves the company.
Our reflex is then: more training, more documentation, more handover sessions. Well-intentioned, but fundamentally inefficient. We make the same investment over and over again, with every new employee and every new team. That is not knowledge management. That is repeatedly reinventing knowledge.
Today, everyone wants to "put AI to work on their knowledge". But as soon as companies take that step, it suddenly becomes clear that:
AI is not a magical solution here. It is a magnifying glass. What is messy becomes messier faster. What is well organised, on the other hand, suddenly becomes incredibly powerful.
In practical terms: an AI assistant drawing from outdated procedures will give outdated answers with just as much confidence as if they were correct. Contradictory documents lead to contradictory AI answers. And without clear ownership, no one knows who is responsible when things go wrong. Garbage in, garbage out. Only faster, more widely distributed and wrapped in a layer of AI authority that makes it sound even more credible.
If we keep saying that knowledge is our product, we need to treat it that way too. Not as a by-product of smart people, but as a strategic asset that requires care, structure and maintenance. Not by locking everything down, but by taking ownership and governance seriously.
Happy employees remain crucial. But without a shared, living knowledge base, every investment in people, and in AI, remains vulnerable.
People can have feet. Knowledge needs roots.
You usually only discover whether your knowledge base is ready for AI when it is already too late. When the chatbot gives the wrong answer, or when the only person who knew the right one has just left. With our Modern Workplace approach, we help organisations tackle exactly that challenge: from structuring your digital workplace to driving the adoption that ensures people actually use that structure.
Curious how mature your knowledge base really is? You won't find out with a quiz, but with a good conversation.
Not by saying "it is neater this way", but by showing what it delivers: faster onboarding, less variation in quality, lower risk when employees leave and AI that actually delivers value. Translate it into a concrete figure: what does it cost today when an experienced employee leaves? That usually lands faster than a long-term vision.
You really shouldn't, at least not without risk. Start with a focused knowledge audit, assign ownership per domain, clean up what is used most and only feed your AI application with information you genuinely trust.
The role shifts from archivist to architect: safeguarding governance, prioritising quality over quantity and bridging the gap between subject-matter experts and systems. That way, knowledge does not only live in people's heads, but effectively becomes part of both the knowledge base and the AI application.
