A Maturity Model for AI Transformation

AI is changing work, value creation, and roles so dramatically that this transformation should not be left to chance but actively shaped. A maturity model can help provide orientation, highlight fields of action, or measure progress in the transformation. But most important from my point of view is having a shared language for the complexity of change in order to jointly develop measures and recognize successes.

AI maturity can rarely be reduced to the question of which tools a team uses. Anyone who looks only at tools easily overlooks the fact that technology without strategy, without competence, and without responsible guardrails often has little impact. That is why I use a model I developed myself, consisting of the following five dimensions.

The five dimensions

Strategy & Portfolio describes the extent to which AI initiatives are intertwined with the goals of the team and the organization. The spectrum ranges from isolated, uncontrolled experiments to the point where AI becomes the central driver of strategy and even sparks new product ideas.

Platform & Tooling looks at access to AI tools and their standardization. This can range from occasionally used, non-approved tools to a company-wide infrastructure that consistently supports AI initiatives.

Delivery & Automation examines how deeply AI is embedded in actual work processes. Here the spectrum runs from isolated manual attempts through firmly embedded workflows to largely automated, AI-supported service delivery.

People & Culture focuses on self-confidence, skills, and the learning culture around AI. It is about the journey from uncertainty and lack of capabilities, through lived curiosity, to AI fluency, where teams proactively rethink the way they work.

Risk & Responsible AI describes governance, policies, and compliance. This ranges from “there are no rules” to responsible principles that are deeply embedded in processes and decisions.

Five levels for orientation

Each dimension can be classified along five maturity levels: Ad-Hoc (isolated experiments without structure), Exploring (initial pilots, growing curiosity), Repeatable (established methods and workflows, initial guardrails), Scaled (cross-team automation, managed portfolio, integrated platforms), and AI First (AI shapes the strategy, high competence, responsible AI is a given).

One point is particularly important to me: the levels are hardly to be understood as an evaluation from “bad” to “good.” There may be good reasons why a team deliberately does not aim for “AI First” in certain dimensions. The model makes areas for action visible while leaving open how a team tackles them in concrete terms. It helps with the questions “Where are we now?” and “Where do we want to go?” The “How do we get there?” remains the responsibility of the team. This openness is part of the design because it invites reflection instead of prescribing recipes.

Simple enough to work directly in a workshop

The real charm of the model emerges when you use it as a coaching tool directly in a workshop. It is simple enough that a team can understand it without lengthy introduction and apply it immediately.

In practice, it often works like this: the five dimensions with their levels are displayed as a matrix on a board, and participants place themselves via dot voting where they see their team today. Within a few minutes, a visible picture of the mood emerges, including the exciting moments when the dots are spread far apart.

So that sticking dots leads to real insight, you need the right reflection questions. For example: What guided you in your assessment? Where do our views diverge more strongly and what could be the reason? Which dimension is slowing us down the most right now? And where would a single step forward help us the most?

Beyond the workshop

What works in the room as dot voting can just as easily be used as an online survey. This way, you can run the measurement across the entire organization and gather insights without elaborate workshops. A follow-up measurement enables a status assessment under the same conditions and shows whether anything has changed.

A maturity model doesn’t take strategy or implementation off anyone’s hands. But it does create the framework that makes both open to negotiation in the first place. In this sense, it works more like a mirror that a team holds up to itself – and that, in itself, can set surprisingly much in motion.

[[ai-hint]]