Know exactly what AI can take on in your most important decision.
We take one high-stakes decision and sort it by who could actually do each part. The parts AI can do, you hand to AI. The part that needs your people’s judgement stays with them, and gets the time it deserves.
A pricing exception. A credit approval. A client recommendation. The calls your business makes again and again, with real money or real risk on them.
AI handles the bottom three. The top is your people’s judgement, and that’s the part to protect.
We trace every part back to where it comes from.
Where a part traces to a source you can point to, AI can draft it and your people check it. Where it traces to nothing written down, you’ve found the judgement. That part is yours, and no AI can reproduce it.
The standards and your own written playbook. They trace to a source, current and checkable. AI can carry these.
What this client is ready for is the part that decides the recommendation. It traces to nothing written. It lives in your people, and that’s where it stays.
See this on your own decision→Activity-based costing, for your knowledge.
Most AI advice is organised the way the technology is sold: by vendor, by platform. The audit starts from the other end. One of your own decisions, and which parts of it AI can take on.
Activity-based costing took one product line and asked what each step costs. The Knowledge Audit asks the same question one level deeper: take one decision and ask what kind of knowledge each part runs on. Then you can rebalance: hand the lower parts to AI, and give your people back the work that grows the business.
Most organisations find the same thing. The work that grows the business is being crowded out by work anyone could do. It doesn’t get done badly. It just never happens.
Activity-based costing made the economics of business activity visible. A Knowledge Audit does the same for your knowledge.
See this on your own decision→Two stages. The first stands on its own.
The audit comes first and is complete on its own. The second stage goes deeper, for the decisions worth it.
Find where your real judgement sits.
We take one decision your business makes again and again, rebuild it from three sources in turn — the public record, your own documents, then the people who make the call — and test every part against where it comes from. Then we draw the line: where the written ground runs out, your people’s judgement takes over.
Make one person’s judgement the team’s.
For the decisions worth it, the second stage takes the judgement that sits with one person and writes it down. Working with the person who makes the call, we turn it into something the whole team can use, traced to its source, and able to stand up when it’s challenged.
One decision. A dry run before you build anything.
The audit is a dry run of AI on a real decision, before any plumbing or build spend. The clearest first step there is.
One decision, rebuilt from three sources and tested part by part, then the line drawn.
About two weeks. No plumbing, no build spend.
A map of where your judgement sits, and proof of what holds up.
It draws the line on one decision. It doesn’t build the tool. That’s stage two, if you want it.
It tests one decision in depth. It isn’t a whole-stack review. The Architecture Audit does that.
It reads what your people know. It doesn’t replace their call. The judgement stays theirs.
The limits are what keep it honest.
Maps the whole stack, and points to the one decision worth going deep on.
You are here. One decision, taken all the way down.
Checks the AI spend you’ve already funded against the map.
Start a conversation.
One decision you can’t get wrong, where the expertise sits in a few people? Tell us which one, and we’ll show you what a dry run looks like. No deck, no pitch.
Start the conversation→Every engagement is led hands-on, and we take on a small number at a time.