MODELCATALOG

Where your organisation is with AI, where it should aim, and what to do next.

AI adoption maturity

A self-assessment in three steps. Describe your organisation, rate 25 capability areas against plain-language statements, and get a maturity estimate with a sequenced roadmap. It takes 15–25 minutes, and nothing you enter is stored.

Start assessment

How it works

  1. Discover. Thirteen questions about your scope, industry, regulation, data and ambition. These set a target level for each dimension. The highest level is not always the right one.
  2. Estimate. For each capability area in scope, pick the statement that best matches how it is owned, planned and funded today.
  3. Plan. Gaps between today and your target become a roadmap in three waves, each item placed by effort against value and linked to related areas worth tackling together.

The five levels

1 Exploratory 2 areas first required here Learning what AI can do here, through limited, contained experiments.
2 Implemented 11 areas first required here Some AI is in production, with basic delivery, data and risk practices behind it.
3 Aligned 10 areas first required here AI is tied to business goals, workflows are being redesigned, and results are measured.
4 Scaled 1 area first required here AI runs reliably across the organisation, on shared standards that units can tailor.
5 Future Ready 1 area first required here AI is part of how the organisation works, and new advances are absorbed predictably.

How it is scored

Each capability area is rated on three indicators: accountability, planning and resourcing. An area scores its weakest indicator, and a dimension scores its weakest area. Your overall level is the highest level whose required areas all reach it.

That is the model's own default rule, and it is strict on purpose: one neglected area holds everything back. The results page also shows the average, so you can see where you are close to the next level.

Structure based on The AI Adoption Maturity Model v1.0, Carnegie Mellon University Software Engineering Institute (SEI) with Accenture, 2026. Descriptions are paraphrased. This is an indicative self-assessment, not an SEI assessment, which relies on interviews, workshops and evidence review by trained assessors.