(This post originally appeared on my LinkedIn Profile and received a lot of attention there, feel free to read the comments there.)
I’ve spent my entire career fighting the same battle.
Someone finds out I work in knowledge management, and within minutes the conversation has shifted to SharePoint, or Confluence, or whatever platform is currently being championed as “the solution.” These days, that platform increasingly has “AI-powered” somewhere in the marketing copy.
And I understand why. Technology is visible. It has a price tag, a vendor, a demo. You can point to it in a procurement document. Knowledge — the human, messy, relational kind — is much harder to put in a slide deck.
But here’s what two decades in KM have taught me: the platform is never the problem. And it’s never the solution either. Neither is the AI layer on top of it.
When a knowledge base goes stale three months after launch, that’s not a technology failure. It’s a culture failure. When frontline staff can’t find what they need in the moment of need, it’s usually not because the search function is broken — it’s because the knowledge was never captured in a way that reflects how those people actually think and work.
Now imagine pointing an AI at that same broken foundation. A large language model is extraordinarily good at generating fluent, confident-sounding responses. It is only as trustworthy as the content it draws from. Outdated articles, undocumented workarounds, tribal knowledge that never made it into the system, contradictory guidance left unresolved for years — AI doesn’t fix any of that. It amplifies it. Confidently.
This focus on people and relationships is the heart of Radical Knowledge Management.
Radical KM starts from a simple premise: knowledge lives in people, not systems. Systems can support its flow, but they cannot create it, sustain it, or replace the trust and relationships that make people willing to share what they know. If you haven’t invested in those human conditions first, you can implement the most sophisticated AI-powered platform on the market and still end up with an expensive, high-velocity misinformation engine.
Moving beyond “KM is a technology problem” requires three honest conversations most organisations are reluctant to have:
Who owns knowledge here — and who is excluded from that? Knowledge governance isn’t about permissions and taxonomies. It’s about power. Whose expertise is captured and valued, and whose is invisible? AI will simply encode and accelerate whatever answer your organisation has already given to that question.
What does your culture reward? If people are measured on individual output and knowledge-sharing is an afterthought, no tool will change that — and no AI will surface knowledge that was never shared in the first place.
What does “good” actually look like for the people doing the work? Not for leadership. Not for auditors. For the person answering their fifteenth call of the day, under pressure, needing to find an accurate answer quickly. AI can be a powerful ally for that person — but only if the knowledge behind it has been curated with their reality in mind.
Technology has an important role. I’ve spent considerable time helping organisations choose and implement platforms, assess readiness, and design governance frameworks. The tools matter. AI genuinely matters. But they’re the last conversation, not the first.
The first conversation is about people. What they know, how they share it, whether they feel safe doing so, and whether the organisation has created the conditions for knowledge to actually flow.
Until we’re willing to have that conversation — really have it — we’ll keep buying new platforms, adding AI on top, and wondering why nothing changes.