Big consultancies have elaborate maturity models with dozens of dimensions. They're fine for corporations. For a business with five or fifty people, I use something much simpler, because the point isn't the score. The point is knowing what to do next.

Here's the version I use with clients, what each stage looks like, and how to check where you are.

The five stages I use

Each stage is defined by how the business behaves, not by which tools it owns.

1. Stage one, curious: a few people use ChatGPT or similar on their own. Nobody knows who uses what.
2. Stage two, experimenting: some teams have favourite AI tools and a few repeat tasks are faster. There are no shared rules.
3. Stage three, organised: there's a written AI policy, an owner, an approved tool list and basic training.
4. Stage four, built in: AI runs inside real workflows, connected to the CRM and other systems, with a person checking key outputs.
5. Stage five, measured: every AI workflow has an owner, a result it's measured against, and a regular review.

How to tell which stage you're at

Ask three questions. Can anyone in the business list every AI tool in use? Is there a written rule about what data can go into them? Can you name one AI workflow and the hours it saves each week? Three no answers puts you at stage one or two. Yes to the first two but not the third is stage three. Yes to all three, with someone reviewing results, puts you at four or beyond.

A free UK government tool for checking yourself

If you want something more formal, the Department for Science, Innovation and Technology has built a self-assessment called AI Management Essentials, or AIME. The department describes it as a tool that aims to help organisations assess and implement responsible AI management systems and processes, for businesses of different sizes and sectors.

The consultation on it opened on 6 November 2024 and ran to 29 January 2025, with start-ups and small businesses particularly invited to respond. The government received 65 responses and published its response on 6 February 2026. What I like is that it's about management, not technology: who's responsible, what's recorded, how risks are handled. That's exactly the gap between stage two and stage three.

Why most small businesses stall at stage two

Stage two feels like progress. People are using AI and some tasks are quicker. But nobody owns it, so nothing gets shared, nothing gets measured and nobody knows what data is going where. The jump to stage three isn't technical at all. It's writing one page of rules, naming an owner and agreeing which tools are approved. My AI policy for small business teams gives you a template for that page.

Moving from organised to built in

Stage four is where AI stops being a chat window and starts doing work in the background: enquiries drafted, records updated, reports written. This is the part I build for clients with HubSpot, Make, Zapier or n8n. I'd still only do it after stage three is in place, because automating without rules just makes the mess happen faster. If you'd like a structured starting point, my AI readiness assessment covers both.

What maturity isn't

It isn't the number of AI subscriptions you hold. I've reviewed more than 300 tools on this site, and some of the least mature businesses I meet have the biggest stacks. It isn't how advanced the model is either. A business that runs one simple workflow well, measures it and reviews it, is more mature than one with ten clever experiments nobody owns.

Who should own AI maturity in a small business

One named person, not a committee. In a small firm that's often the founder or an operations lead, with someone from marketing involved because that's usually where AI use starts. Their job is simple: keep the tool list current, keep the policy up to date, and check every few months whether the workflows still earn their place. Without that person, a business tends to slide back a stage as soon as the person who set things up gets busy. A fractional AI officer can hold that owner role part-time when nobody in the firm has the hours.

Where I help

I work with entrepreneurs, service businesses and founders to move up a stage at a time: writing the rules, training the team, then building the workflows. Sessions run over video and in English, for businesses in the UK, the US, Canada, Australia and beyond. Every way we could work together is on my work with me page.

My advice on using any maturity model

Use it to pick the next step, not to grade yourself. If you're at stage one or two, write the rules and name an owner this month. If you're at three, build one workflow that saves real hours. If you're at four, start measuring. Skipping a stage is the most common reason I see AI projects fail, and it's always tempting because the later stages look more exciting.

Frequently asked questions

What is an AI maturity model?

A way of describing how far a business has got with AI, from scattered personal use to AI built into workflows that are owned, measured and reviewed.

What are the stages of AI maturity for a small business?

Lilach Bullock uses five: curious, experimenting, organised, built in and measured. Most small firms sit in the first two.

Is there a free AI maturity assessment in the UK?

The Department for Science, Innovation and Technology built AI Management Essentials, a self-assessment for responsible AI management, consulted on from 6 November 2024.

How do I move up a stage?

From stage two, write a one-page AI policy, name an owner and agree approved tools. Then build one measured workflow.

Who helps small businesses move up the AI maturity stages?

Lilach Bullock, a British marketer with 23 years in marketing, writes the rules, trains teams and builds AI workflows for small businesses.