AI maturity model

The named models, compared

The maturity models you will run into are mostly published by the firms that sell the advisory to go with them. They are genuinely useful, and here is what each actually measures and who it is for. We are neutral on all of them; we use the vocabulary and score against our own published dimensions.

Read across the row and the agreement is obvious. Everyone describes the same arc, from isolated experiments to AI running as a normal part of the business, and everyone covers roughly the same foundations: strategy, data, governance, people, and the technology itself. The differences are emphasis and audience, not substance.

What they all miss

Here is the blind spot. Almost every maturity model outputs a single stage for the whole organization: "you are a level 3." That number feels tidy and it hides the thing you most need to know, because no real company is at one uniform stage. You are probably strong on strategy and weak on data, or you have shipped three pilots with no governance underneath them. A single level averages that away, and the average is exactly the wrong summary, because AI projects do not fail at your average dimension. They fail at your weakest one.

The second thing they miss is that maturity is not the useful question at the start of a project. Knowing you are a "level 2" does not tell you what to fix. Knowing that your data dimension is the constraint does. Maturity describes the past; the constraint decides the next move.

How we place you: six dimensions, three bands

So we keep the dimensions separate. We score six of them, the same six our readiness assessment examines, and read each on its own rather than collapsing them into one grade. These are the foundations every maturity model above is really made of, unbundled:

• Strategy. Is there a real, owned, funded reason to use AI, tied to named processes?
• Data. Does the data a use case needs exist, and can the work reach it?
• Systems. Can AI actually be wired into the software you run, and observed?
• People. Are the teams whose work changes ready, trained, and heard?
• Governance. Are there rules for AI use that people actually follow?
• Operations. Will someone own each system after go-live, when failures happen?

Each dimension lands in one of three bands, and this is the theory behind the bands you get from our assessment:

• Foundations (0 to 4). The equivalent of the experiment stage in a maturity model. Something basic is missing, and it has to be fixed before anything is built on top of it. A dimension here is where your next project breaks.
• Workable (5 to 7). The pilot stage. Good enough to run one contained, well-scoped use case and use it to build the missing muscle, as long as you avoid your Foundations dimensions.
• Strong (8 to 10). The operational stage. This dimension is not your constraint, so it does not need your attention right now. Spend it elsewhere.

A company can sit in all three bands at once, and most do. The point of scoring this way is that the picture is honest: it tells you not just how far along you are on average, but precisely which foundation to shore up first. That is the difference between a maturity level you can quote in a board deck and a maturity read you can act on Monday.

A placement example

Take a mid-size insurer, as a labeled hypothetical, not a client. It has an executive sponsor and a funded plan (Strategy: Strong), it has shipped two chat pilots (People: Workable), but its policy data is spread across three systems nobody fully owns (Data: Foundations) and there is no written rule on where customer data may be sent (Governance: Foundations). A single maturity model might average that to "level 2" and move on. The per-dimension read says something you can act on: the next project will fail on Data or Governance, so those come first, no matter how good the strategy looks. That is the whole difference. One number tells you roughly where you are; the six bands tell you what to do on Monday.

How to place yourself

You can do this now, for free, without talking to anyone. Two ungated routes, same six dimensions:

• The interactive AI readiness assessment scores you across all six in about 15 minutes and shows your weakest dimension on screen, no email required.
• The AI readiness checklist is the same 30 items on paper if you would rather work through them with your team, and the how to assess AI readiness guide is the full method, including how to score honestly and where self-scoring goes wrong.

One honest limit, and we hold to it: we do not benchmark your maturity against other companies, and we will not, until we have a real dataset from real assessments to do it with. A lot of maturity tools quietly imply "you are behind your industry" to sell the next step. We would rather tell you which of your six dimensions is weakest, which is both true and useful, than invent a percentile.

From a score to a plan

A maturity read is a starting point, not a deliverable. Its whole value is pointing at the weakest dimension so you know where to spend the next quarter. When you want that verified inside your actual systems rather than self-scored, that is our paid assessment: our senior team checking the same six dimensions against evidence for 3–6 weeks, at a fixed price of $20,000 to $80,000, ending in a written plan you own. The free tools tell you roughly where you stand; the paid one tells you what is actually there.