The frameworks
How I think
about companies.
Three connected ideas. The first says what a company should be able to do. The second says how intelligence compounds. The third says where each kind of intelligence belongs.
Before the frameworks
Here's how I actually see a company.
From the inside, this is what it looks like.
Everything connected to everything, and none of it labelled.
Most people respond by working harder inside their own piece of it.
My job is making it legible — so the work moves without a person carrying it.
And that's what these three ideas are for
Three frameworks, in the order I actually use them.
01 — Operating Intelligence
A company should be able to answer five questions without calling six people into a meeting.
Most companies can answer pieces of it. Very few can answer all of it continuously. That gap is why decisions arrive late, work gets duplicated, customers get bounced around, and leaders end up managing by anecdote.
What is happening?
Signal from every part of the business, in one place, without asking anyone to assemble it.
Why is it happening?
Context attached to the signal, so the number means something before someone argues about it.
What should happen next?
A decision with an owner and a deadline, not a discussion that recurs monthly.
Did it actually happen?
Execution observed, not assumed. Most organizations discover the answer a quarter late.
What did we learn?
The outcome changes the system. Otherwise you relearn the same lesson with new people.
A business that can do all five, continuously, is what I mean by Operating Intelligence. Not another dashboard, not another automation, not another agent — the intelligence is in the connection between them.
02 — The doctrine
AI² — augmented intelligence.
Once the business is visible as one system, AI becomes a much better question. Not "what can we automate?" but "what should machines carry, so people can do more of what only people are good at?"
Machines are extraordinary at
- Memory
- Search
- Comparison
- Monitoring
- Routing
- Summarizing
- Repetition at volume
People are extraordinary at
- Judgment
- Persuasion
- Trust
- Leadership
- Reading nuance
- Taking responsibility
- Deciding what matters
The mistake is asking one to do the other's job. The goal isn't a company with the most AI. It's a company that uses every form of intelligence where it creates the most value.
03 — The allocation framework
The Human Gradient.
AI isn't a replacement decision. It's an allocation decision.
There's no universal answer. A payroll reconciliation shouldn't have the same human-to-AI mix as firing someone. Every task sits somewhere on this gradient — and the cost of being on the wrong side of it is different in each direction.
The goal isn't maximum automation. It's the right intelligence, in the right place, at the right moment.
Keep going