Maturity without adoption is worthless. Many organisations are not "immature". They are stalled: governance, platforms and policies are in place, but few people use AI in their daily work. Measure maturity and adoption separately.
Short answer: AI maturity describes how well an organisation has the foundations to create value with AI: strategy, data, technology, skills and governance. But high maturity says nothing about whether employees actually use AI. To know where you stand and what to do next, you need to look at both.
What is AI maturity?
AI Sweden, Sweden's national centre for applied AI, describes AI maturity as an organisation's capability to use AI technology and applications, taking into account factors including strategy, data infrastructure, technology platforms, people and governance (AI Sweden, AI maturity assessment, in Swedish).
Its maturity ladder, developed with appliedAI in Germany, has five steps: Beginner, Explorer, Practitioner, Professional and Leader. Each step is defined by how much value the organisation has created with AI (AI Sweden, maturity ladder, in Swedish). AI Sweden stresses that the top step isn't necessarily the goal for everyone, and that the model is meant to help understand the current state, not to judge.
It is a good model, and widely used in Sweden. Similar models exist across consultancies and vendors worldwide.
What most maturity assessments miss
Most AI maturity assessments ask about foundations. Is there a strategy? Is the data in order? Are there platforms, policies, skills? The answers place the organisation on a single low-to-high scale.
The problem is that foundations and usage don't always move together. We see two common patterns:
Strong foundations, low usage. The strategy is written, licences bought, policy approved, and daily work looks the same as before.
High usage, weak foundations. Employees use AI heavily, often in their own tools, with no shared rules, data or follow-up.
Both can land on the same point of a maturity scale. They need completely different next steps.
Two axes instead of one
At Lyon AM we place organisations on a matrix with two axes: adoption (how much and how well AI is used in daily work) and maturity (how well the foundations are in place). That gives four positions:
| | Low maturity | High maturity | |---|---|---| | High adoption | Scattered: lots of experimentation, little governance | Scaled: the goal, AI is part of how work gets done | | Low adoption | Emerging: the start of the journey | Stalled: everything in place, few people using it |
Each position needs something different:
Emerging: Start small. Pick a few workflows, have leadership go first, set simple rules.
Scattered: Capture what already works. Offer approved tools, shared rules and a way to spread good practice, without killing the energy.
Stalled: Stop investing in more foundations. The problem is leadership, expectations and ways of working. It is change work, not a technology project.
Scaled: Keep going. Follow up, renew, and make sure new tools and new joiners enter the same rhythm.
Why "stalled" is so common
Ask leadership teams how AI is going and they often answer with foundations: we have a strategy, a platform, a policy. Those are easy to measure and easy to report. Adoption is harder to measure, and less flattering.
Prosci's study of 1,107 professionals found 63% of AI implementation challenges are human factors and 16% technical (Prosci). At Lyon AM our view is that roughly 70% of a successful AI programme is about people, leadership and ways of working. A maturity assessment that mostly measures technology and governance therefore only sees the smaller part of the picture.
Finding out where you stand, in four steps
1. Measure adoption on its own
What share of employees use AI in their work each week? For what? Have workflows actually changed? Licences and logins are not enough.
2. Measure foundations on their own
Strategy and priorities, leadership commitment, data and tools, rules and governance, skills. Established models such as AI Sweden's ladder fit well here.
3. Place yourself on the matrix
Combine the two. Be honest: high maturity with low adoption is not success.
4. Choose the next step by position
Don't apply the same plan to every position. A stalled organisation doesn't need a new strategy. It needs managers to start using and asking for AI.
Frequently asked questions
What is AI maturity? An organisation's capability to create value with AI, considering strategy, data, technology, people and governance.
What is an AI maturity model? A framework that describes stages of AI capability. AI Sweden's has five: Beginner, Explorer, Practitioner, Professional and Leader.
How do you assess AI maturity? Usually with a survey on foundations, followed by analysis and a workshop. Add measures of actual adoption in daily work.
Does every company need to reach the highest maturity level? No. Even AI Sweden stresses it is a strategic choice, and for many organisations a lower target is enough.
Next step
Find out where you stand on both axes. The Lyon AM AI diagnostic places you on the matrix and gives a first recommendation for what to do next. Want to go deeper? Contact us.
