You have read about tacit and explicit knowledge. You understand that the things your people know — the judgment, the pattern recognition, the hard-won operational intelligence — are more commercially valuable than any policy document or template. You understand that AI commoditised explicit knowledge and that tacit knowledge is now the only sustainable advantage.
The question is practical: what do you actually do about it?
Start with the people, not the technology
The instinct is to start with the platform. Sign up, log in, start building. That instinct is wrong — or at least, premature.
Tacit knowledge lives in people. Before you can capture it, you need to know where it sits. And in most firms, it sits in three or four people who have been there longest, who know the most, and who have never been asked to articulate what they know.
Start there. Not with a formal knowledge audit. Not with a consultancy engagement. Just a conversation: what do you know that nobody else in this firm knows? What would we lose if you left tomorrow? What do you do differently that the new hires have not figured out yet?
The answers will surprise you. Not because the knowledge is exotic, but because it is so embedded in daily practice that the people who hold it do not recognise it as special. They just think of it as doing the job.
Capture the judgment, not the process
The biggest mistake firms make when trying to capture knowledge is focusing on process documentation. Step one, do this. Step two, do that. That is explicit knowledge. It is useful but it is not what differentiates you.
The tacit knowledge is in the judgment calls. Why step two sometimes gets skipped. Why the process says to do X but experienced people do Y instead. Why certain clients get a different approach. Why the deadline in the system is not the real deadline.
When you capture knowledge, focus on the exceptions, the workarounds, the "we learned this the hard way" insights. Those are the things AI cannot get from public data. Those are the things that make your firm's output different from the generic version.
Make it easy to contribute
If capturing knowledge feels like writing a manual, nobody will do it. People are busy. They have client work. The moment knowledge capture becomes a separate task — something added to their day rather than woven into it — participation drops to zero.
The mechanism matters. Knowledge hubs are designed to accept contributions in the way people naturally communicate: short observations, specific insights, practical guidance. Not essays. Not formal documentation. Just the things people know, captured in the way they would explain them to a colleague.
"HMRC always queries R&D claims over 50k in the first year. Build in two weeks for the enquiry."
"Client X goes quiet in Q4 every year. Not a churn signal. They are just busy with year-end."
"The standard template says 14 days. Give them 21. The relationship is worth more than the clause."
That is tacit knowledge. Three sentences. Each one more valuable than a hundred pages of policy documentation.
The constitution tells AI how to think. Knowledge hubs tell it what to know.
Two systems work together. The constitution captures your firm's voice, values, standards, and guardrails — the how. Knowledge hubs capture your firm's domain expertise, client intelligence, and operational reality — the what.
Without a constitution, AI has no identity. It defaults to generic. Without knowledge hubs, AI has no depth. It defaults to public knowledge.
With both, AI produces output that sounds like it came from your best person on their best day. Because it is drawing on the same knowledge your best person uses — structured, accessible, and available every time.
Start small. Compound fast.
You do not need to capture everything on day one. You do not need a complete knowledge base before AI becomes useful. Start with one area. One client type. One process where you know the experienced people do it differently to what the manual says.
Capture five insights. Then ten. Then twenty. Each one makes AI slightly better at doing things your way instead of the generic way. The improvement compounds. By the time you have a hundred contributions across your team, the difference between your AI output and a competitor's is visible in every interaction.
The firms that start now build an advantage that widens every week. The firms that wait will eventually start from the same place — but the firms that started earlier will be a thousand contributions ahead.
The practical path
Step one: identify the two or three people in your firm who hold the most tacit knowledge. They are usually senior, usually long-tenured, and usually the people everyone goes to with questions.
Step two: have the conversation. What do they know that is not written down? What would the firm lose if they left? What shortcuts, exceptions, and judgment calls do they make that newer staff do not know about?
Step three: start capturing. Use the knowledge hub. Short contributions. Specific insights. Do not aim for comprehensive — aim for useful.
Step four: use it. Ask AI to do a task it has done before. Compare the output with and without the knowledge hub. The difference is immediate and obvious.
Step five: keep going. Make contribution a habit, not a project. The more knowledge flows in, the more valuable every AI interaction becomes.
That is it. No transformation programme. No twelve-month roadmap. Just start capturing what your people know, and let AI use it.
See how the platform captures and applies your firm's knowledge. →