We've spent twenty years building the systems underpinning these firms' knowledge. We know what has to be structured before AI can help. We know what isn't worth touching.
Get in touchThe report that gets rebuilt from scratch every time because the good version is buried in someone's old files. The way you actually run a project, which only works because the people who know are in the room, and falls apart when they're not. The standard everyone follows and no one has written down. This is the knowledge that runs the firm, and almost none of it sits anywhere that software can reach.
So when AI arrives, it works on the thin layer that is written down and misses everything that matters. That's not a model problem. It's a knowledge architecture problem, and it's fixable.
Not by buying more tools. Every firm has real work an AI can help with: administration, document preparation, recurring bottlenecks, decision support. The question is: what has to be structured first, what should remain human judgement, and what simply isn't worth automating. The answer to this question is the hard part.
?
The question is: what has to be structured first, what should remain human judgement, and what simply isn't worth automating.
We help professional services firms work out where AI can help, where it can't, and what knowledge and judgement sit underneath both.
Our work is practical and bounded. We start with how the firm actually operates: where knowledge lives, how decisions get made, what information people rely on, and what happens when key people are unavailable.
That means talking to people at every level, not just the people who commissioned the work. The person who runs a process every day usually knows better than anyone which parts are worth automating and which parts quietly depend on judgement. Often the most useful insight comes from the most junior person in the room, because they feel the friction the directors stopped noticing years ago.
This only works if people experience it as support, not extraction. Too often, knowledge capture has implied turning what someone knows into a system that makes them replaceable. We want the opposite: expertise that is easier to find and reuse, and time won back from the work that drains judgement rather than uses it.
From there we help firms:
Sometimes the most valuable recommendation is not what to automate, but what to leave alone.
A short, practical assessment of where AI can genuinely help your firm, and what has to be fixed first.
Typically this gives you:
A solo practitioner has no one to push back on their document before it goes out. The AI that helped write it will not volunteer the criticism. Roundtable is a panel of built-in critics that ask the hard questions, and leaves the judgment to you.
Read more
Not all knowledge behaves the same way when you try to formalise it. This post maps three categories of professional knowledge - and argues that protecting tacit expertise requires deliberate choices about what you leave out.
Read more
Ask most consulting firms whether AI is paying off and the answer is awkward. The tools are everywhere, the results are not. The problem is not the technology. It is the absence of written context - decision rationale, process knowledge, client understanding - the things an agent needs but almost nobody records.
Read moreWe work with a small number of professional services firms at a time.
If you're trying to work out where AI can genuinely help, what foundations need attention first, or whether you're solving the right problem at all, start a conversation.
Start a conversation. Worst case, you get an outside read on where AI fits in your organisation and where it doesn't.