An operating discipline
How work is structured, tooling is arranged, and decisions are recorded, so AI augmentation compounds across the organisation instead of fragmenting into private, disconnected gains.
Instructed Intelligence
The team layer for enterprise AI.
AI makes every individual faster. The organisation is not following. What the machine learns about your business is lost when the session ends, so teams drift apart and the same knowledge is taught again every day. Instructed Intelligence is the methodology and software that closes the loop across a whole organisation, for people and machines alike.
01 Thesis
Everyone is faster. Nothing is quicker.
Each person with an AI is measurably quicker, and each session really is faster. But what they teach the machine is written off when it ends, and taught again tomorrow. That is knowledge debt: paid in time, per head, per session, and booked on no ledger. The individual gain is real. The organisation quietly borrows against it.
02 What we do
Methodology
Advisory
The changing mandate
Leadership's job is shifting from ensuring execution to ensuring the right thing is executed. We help make that shift deliberate, at the altitude it needs.
Substrate
The shared record
How the business speaks, decides and operates, kept as one record that is shared, structured and attributable. Auditable by design; yours to hold.
On-prem · your cloud · SaaS
03 The broken loop
Every AI reads from everywhere. And writes nowhere.
The industry has spent three years optimising the feed: more context, more retrieval, more of the pile. It has not shipped the return. What the machine learns is destroyed when the session ends, and the next one starts from nothing. The fix is not more data, and not a better model. It is the right data: shared, structured, and traceable to something you wrote and stand behind.
04 The silos return
The silos we spent twenty years removing are reassembling. In an afternoon.
Two people, each with an AI, each with their own private context, diverge within a week, and neither can see it, because no shared surface holds either view. It has every property of the silos we dismantled: private, invisible, lost when the holder leaves. And one new one: it forms in an afternoon. Nobody bought an assistant to rebuild silos. It is simply what a tool does when it reads from everywhere and writes nowhere.
05 Human in the loop, or not
The same context aligns every AI on the problem. Not only the people using it.
More and more work runs with no human in the loop, and the human was the error-correcting layer nobody priced. Take it out and a plausible, wrong answer is acted on at machine rate, then becomes the input to the next thing. Coordination does not stop mattering when the participants become machines: systems that read from everywhere and write nowhere cannot align, and the fragmentation hardens into the architecture. One shared, attributable record is what holds people and machines to the same truth.
06 The shift
The high-leverage work is no longer doing. It is making sure the organisation does the right thing, together.
07 Engage
We work with a small number of enterprises and partners at a time.
Start a conversation →