Estimated reading time: 8 minutes

Key points:

  • Adopting AI is a journey, not a switch you flip. Our AI maturity ladder, based on existing models and tailored to Dutch SMEs, has five phases: Exploration, Operational, Scaling, AI-first and Self-learning.
  • Knowing which phase you are in gives direction: you see what the logical next step is and where the quickest win lies.
  • Without that picture you invest at random, a training here and a tool there, without knowing whether it fits where you stand.
  • The top phase is deliberately independence: an organisation that keeps building with AI on its own, without being tied to one supplier, and where the systems themselves keep learning too.
  • You determine your phase with a short scan across a few dimensions, and you measure again later, so you can see your progress in black and white.

Table of contents

  1. Why knowing where you stand matters
  2. The five phases of AI maturity
  3. One question that helps at every phase
  4. How to determine where you stand
  5. What the ladder gives you
  6. How we can help
  7. Frequently asked questions
  8. Sources

Adopting AI is a journey, not a switch you flip. One organisation is still running isolated trials, another is already building new solutions on its own. Knowing where you stand determines what a sensible next step is, and where your time and money are best spent.

For that we use our own maturity ladder with five phases. We developed it based on existing international readiness models and tailored it to the Dutch SME, so the phases are recognisable for the businesses we work with. Below we walk through the five phases, explain how to determine where you stand, and what the next step delivers.

Why knowing where you stand matters

Not knowing where you stand costs you in three ways.

You miss direction. Isolated experiments rarely lead to results, and the knowledge sits with a single person who can just as easily leave the organisation. You invest at random: a training here, a tool there, without knowing whether it fits where you really stand. And you cannot see what it delivers, because without a baseline measurement at the start you cannot prove progress later.

Insight into your phase solves that. It gives direction, it makes your investments targeted, and it makes the return visible. That turns the next step into a choice instead of a gamble.

The five phases of AI maturity

The journey runs from isolated experiments to an organisation that keeps building with AI on its own. Five phases, each with its own picture, its own signals and its own next step.

Phase 1, Exploration. AI consists of isolated experiments. A few people try ChatGPT, without direction. There is no AI vision or owner, usage sits with individuals, and there are no agreements about safe use. The risk of getting stuck: it stays at the level of toys while others push ahead, and the knowledge disappears with the person who started it. The next step is to lay a shared foundation and pick one concrete process to start with. How to go about that you can read in Where do you start with AI in your organisation?, and how to pick the right process in Which processes are suitable for AI?.

Phase 2, Operational. AI helps with fixed, well-defined tasks. Part of the team uses it independently, with separate but deliberately chosen tools and informal agreements. It works, but in a fragmented way, and the gain hangs on individuals rather than on the team. The next step is to connect those separate tasks into coherent processes and bring more people along. How to make individual employees productive with AI you can read in our guide AI as a work partner: how to use AI in your daily work.

Phase 3, Scaling. AI sits in several processes, with the first agreements and measurable gains. AI is in the plans, most people use it, and there is documented policy including the basics of the AI Act. The risk is that growth stalls without a solid data foundation and clear governance: it becomes unmanageable and hard to trust. The next step is to get data and governance in order, broaden the automation, and make managers owners of adoption.

Phase 4, AI-first. AI is the starting point for new work, with demonstrable value. Here the question shifts: from "how do we make this existing process faster with AI?" to "what would this process look like if we set it up again today, with AI as the starting point?". AI sits at the core of value delivery, people build their own ways of working, and governance is embedded. The risk is dependence on suppliers and black boxes: you give up control and margin if you cannot build it yourself. The next step is to become independent, able to build, assess and develop further on your own.

Phase 5, Self-learning. The organisation learns and keeps building on its own, without being tied to one supplier. AI is anchored in the strategy, there is an in-house learning culture where people train each other, and the AI return is part of how the business is steered. At this rung not only the people learn: the systems themselves learn along. They get feedback from use, are adjusted, and become better over time as a result. The risk is standing still, because staying ahead demands continuous learning, from your people and from your systems. The next step is to secure that learning culture and keep pace with the latest developments.

This top phase, independence without lock-in, is deliberately the end point. Most suppliers work towards dependence; we work towards an organisation that can do it itself.

Wireframe diptych: on the left a messy process merely sped up, on the right the same process rebuilt with AI as the starting point.

One question that helps at every phase

The strongest AI applications rarely come from bolting AI on top of something. They come from asking one question: what if we set this process up again today, with AI as the starting point?

In phase 4 that is the normal way of working, but the question is useful at every rung, even if you are still in Exploration. It forces you to break away from "this is just how we do it". Speeding up a cumbersome process a little delivers something; rethinking that same process with AI often delivers much more. You do not have to build the answer right away, just the thought alone sharpens where to aim your first or next step.

How to determine where you stand

You determine your phase with our AI maturity scan: a short questionnaire that looks across a few dimensions at once, because an organisation can be further along on one point than on another. The scan looks at your strategy, your people, your processes, your data and technology, how responsibly you set it up, and how independent you already are.

The outcome places you in one of the five phases and shows, per dimension, where the biggest opportunity lies, usually the weakest point where the quickest win sits. After that it is a simple cycle: measure where you stand, choose what the next step is, take that step in a focused way, and measure again. That way progress becomes a fact you can back up, including towards your board.

Where you are working towards belongs in your AI direction, in which you also set down which processes go first and who leads them. How to draw up that direction you can read in Building an AI strategy: from isolated experiments to a direction.

Wireframe cycle of four stations connecting measure, choose, act and measure again.

What the ladder gives you

Knowing where you stand delivers four things.

  • Direction. One clear picture you can share internally, instead of isolated experiments without coherence.

  • Targeted investments. Every euro goes to the step that fits your phase, not to a random tool or training.

  • Visible progress. The measurement at the start and later makes growth provable and discussable.

  • Independence. The goal is an organisation that can do it itself. That is deliberately the top phase.

A solution built at the top of the ladder shows where the journey leads. The vehicle valuation we built for Bolsenbroek saved more than 1,000 hours and over 65,000 euros in the first year. That is the kind of value that emerges when AI comes to sit at the core of the work.

How we can help

If you want to know which phase your organisation is in and what the logical next step is, we map that out together. In the workshop AI Strategy for Managers we place your organisation on the ladder, discuss where the quickest win lies, and set the first direction.

If you then also want to take that next step, from choice to working application, we pick that up in AI in Business. We measure where you begin, focus on what is needed, and measure again, so the progress stays visible.

Schedule a no-obligation conversation, and we will look together at where the first win lies for your organisation.

Frequently asked questions

What is an AI maturity model?

A model that shows how far an organisation has come with AI, from isolated experiments to building on its own. Our ladder has five phases: Exploration, Operational, Scaling, AI-first and Self-learning. We developed it based on existing international readiness models and tailored it to the Dutch SME. It gives direction and makes the next step concrete.

How do you know where your organisation stands with AI?

With a short scan across a few dimensions: strategy, people, processes, data and technology, responsible use and independence. The outcome places you in a phase and shows where the biggest opportunity lies.

What are the phases of AI maturity?

Exploration (isolated experiments), Operational (AI for fixed tasks), Scaling (AI in several processes with first governance), AI-first (AI as the starting point with measurable value) and Self-learning (building on your own without lock-in).

Does every organisation need to reach the highest phase?

Not necessarily. The right phase depends on your ambition and your work. More important than getting as high as possible is taking a logical next step that fits where you stand now.

How do you make progress visible?

By measuring at the start and again later. The difference between the two measurements shows the growth of your organisation and your team, in a picture you can share internally.

Sources

The five phases form an And AI model of our own. It is inspired by, and broadly comparable to, public readiness frameworks, without being identical to them.