Free resource

AI readiness checklist for small and mid-sized businesses

20 questions. About 10 minutes. No sign-up needed.

Most small businesses already use AI, often without anyone deciding to. A salesperson drafts emails with ChatGPT, the accounting software adds an “AI assistant”, someone translates a contract with DeepL. None of this is a problem in itself. The problem is not knowing it is happening.

This checklist is a quick way to see where you stand. Answer each question with Yes (2 points), Partly (1 point) or No (0 points). Your score updates at the bottom of the page.

1. Know what you use

1. Do you have a list of the AI tools used in the company?

Include the free ones people use on their own: ChatGPT, Copilot, Gemini, DeepL, Canva, meeting transcription apps.

2. Do you know which of your business software has added AI features?

CRM, accounting, HR, helpdesk and office suites have been adding them, often switched on by default.

3. For each tool, do you know who uses it and for what?

“Marketing, for first drafts of social posts” is enough. “Everyone, for everything” is not.

4. Do you know whether you only use AI, or also sell or put your name on something that contains it?

The EU AI Act treats the two very differently. Most small businesses are users (“deployers”). If you build AI into a product or service you sell, you may have provider obligations.

2. Data and confidentiality

5. Have you told your staff what must never be pasted into a public AI tool?

Typically: client data, personal data, contracts, pricing, source code, anything under NDA.

6. For the tools you rely on, do you know whether your data is used to train the vendor’s models?

Free and consumer versions often allow it by default. Business plans usually don’t, but check the settings.

7. Do you know where the data you put into these tools is processed and stored?

Inside or outside the EU matters for GDPR, and some of your clients will ask.

8. If personal data goes through an AI tool, is that use recorded in your GDPR register?

An AI tool that processes personal data is a processing activity like any other.

3. Rules and responsibilities

9. Do you have a written AI use policy, even one page?

What’s allowed, what isn’t, who to ask. Short and clear beats long and ignored.

10. Is one person named as responsible for AI questions?

It doesn’t need to be a full-time role. It does need a name.

11. Can staff easily ask before trying a new AI tool?

If the answer is “nobody to ask”, people will simply try it.

12. Do your client contracts or terms say anything about your use of AI?

Some clients now include AI clauses in their purchasing terms. It’s better to know before they ask.

4. Risks and the AI Act

13. Do you use AI to screen candidates, evaluate staff, or decide on credit, insurance or access to services?

These uses are “high-risk” under the AI Act. Under the current timetable, the obligations for them apply from 2 December 2027. (Answer Yes if you don’t use AI for any of these, or if you do and have already looked at what it implies.)

14. If customers talk to a chatbot or see AI-generated content from you, do you tell them?

The AI Act’s transparency rules have applied since 2 August 2026.

15. Is AI output checked by a person before it goes to a client or drives a decision?

AI tools are confidently wrong from time to time. Someone has to own the final version.

16. Do you know what to do when an AI tool gets something badly wrong?

Who is told, how it’s corrected, whether a client needs to be informed.

5. People and skills

17. Have the people who use AI had at least a short introduction to what it does well and badly?

The AI Act expects organisations to support the AI literacy of the people using these tools. An hour goes a long way.

18. Do managers know which tasks in their teams are now done with AI?

It affects quality checks, workload planning and, sometimes, what you tell clients.

19. Do you review your AI tools and rules at least once a year?

Tools change, prices change, and the rules are still moving.

20. If a client, your insurer or your bank asked how you use AI, could you answer in writing?

This question is arriving in tenders, supplier questionnaires and insurance renewals.

Your score 0 / 40 (0/20 questions answered)

What your score means

30 to 40: a solid base. You know what’s being used and you have some rules. The next step is usually to formalise what you already do and keep it up to date.

16 to 29: some useful gaps to close. Typically the tools are known but the rules, responsibilities or data settings aren’t written down. A few hours of work fixes most of it.

0 to 15: start with the basics. That’s common, and it’s quick to improve. List the tools in use (question 1), agree what must never be pasted into them (question 5), and name one person responsible (question 10). Those three steps cover most of the risk.

Get the AI Act key dates

A page with the deadlines that matter for a small business, then an email when the timetable changes. No monthly newsletter.

We'll send you a confirmation email. Privacy policy

Want to go through your answers with someone? A first conversation is free and without commitment.

Let’s talk

Dates are based on the EU AI Act as amended by the Digital Omnibus, as of November 2026. This checklist is general information, not legal advice. Your answers stay in your browser: nothing is sent or stored.

Sources