11 February 2026 · AI Strategy · 2 min

The AI Readiness Framework for Operators

Most AI readiness assessments measure whether you have the tools. The question that predicts anything is whether your decisions are written down.

Every AI readiness assessment I have been handed measures the same four things: data, tooling, talent, and something called culture. None of them predict whether a company will get anything out of this.

Here is the one that does. Are your decisions written down, and is the reason attached to them?

Why that question and not the others

A model is a very fast reader with no memory of your company. Everything it can help with is a function of what it can read. A company whose decisions live in people's heads and in Slack threads from 2023 has nothing to hand it, and no amount of tooling budget changes that.

This is also why the companies that do best with AI are often the boring ones, the ones with a real changelog, a real decision log, a brand system that is data rather than a PDF.

The work is not adopting AI. The work is becoming legible to it, and most of that work was worth doing anyway.

The four questions, in order

  1. Can a new hire find the current answer to a common question in under a minute? If not, neither can a model.
  2. When a decision is reversed, is the reversal recorded next to the original? Reasons are the part that does not survive summarization.
  3. Who decides when a new question appears? A process with no named decider produces a backlog, not a decision.
  4. What are you prepared to stop doing? Adoption without subtraction is just more surface area.

What this is not

It is not a maturity model, because maturity models let you score well by buying things. Every question above is answered by writing something down, which is cheap and unglamorous and the reason nobody sells it to you.

Start with the second question. It is the one most companies fail and the one with the shortest path to passing.

Next note

Why Positioning Is a Translation Problem