AI tools for small business: where they help and what to watch for

A lot has been said about AI tools and little of it sober. This article tries to be the latter. We describe what Dutch companies do with AI according to Statistics Netherlands (CBS), what the European AI Act asks of you, and how to start with a small, manageable step. We do not recommend tools by name and do not promise time savings or revenue: those depend on your work and on how carefully you set it up. All facts have a source at the bottom.

What do other companies do with AI?

At the end of 2025 the CBS published figures on AI use by companies. According to the CBS, 33 percent of companies with 10 or more employees used AI in 2025; in 2024 it was 23 percent and in 2023 14 percent. The CBS calls the figures provisional. Text mining was used by 22 percent and natural language generation by 16 percent; 10 percent had considered AI but not used it. Companies used AI most often for marketing or sales. For smaller companies the percentages are lower: the CBS gives 17 percent for companies with 2 or more employees and 14 percent for companies with 2 to 10 employees.

What do we learn from that? That AI is no longer a niche, that marketing and sales are the most popular use, and that most small businesses do not (yet) work with it. That does not mean you have to. It means you can choose calmly.

Where do AI tools usually help?

Based on what the CBS describes, use lies mostly in text: developing ideas, structuring texts, summarising and writing drafts. The following are our own practical rules of thumb (no figures, no guarantees):

• Drafts and structure. A first outline of a quote letter or newsletter, which you review yourself.
• Summarising. Condensing long documents or emails, with you checking that it is right.
• Rewriting. A text shorter, clearer or in a different tone.
• Translating and proofing. A first translation, after which someone who masters the language checks it.
• Routine work. With automation you connect tools so data does not have to be retyped by hand; see automation for small business.

What to watch for

A language model writes convincingly, even when it is wrong. Google says generative AI can be useful for research and for adding structure to original content, but that generating many pages without added value may violate its spam policies. So check every fact and let nothing leave unread. More about the approach for your website in content strategy for small businesses.

Privacy: what do you enter?

As soon as you put data about customers or employees into an AI tool, you are dealing with personal data, and the GDPR applies. It asks, among other things, for a legal basis, transparency and appropriate security. Practical questions: where is the data processed, does the provider use it for training, is there a data processing agreement, and can you have data deleted? If you cannot find the answer, do not enter personal data. See also cookies and GDPR for your website and the basics of website security. The NCSC publishes general security advice for organisations.

What does the European AI Act ask?

The AI Act (Regulation (EU) 2024/1689) has been in force since 1 August 2024. According to the European Commission, prohibited AI practices and the AI literacy obligations have applied since 2 February 2025, and the regulation has largely applied since 2 August 2026, with exceptions and extended transition periods for high-risk uses. From 2 August 2026 the AI Office and the authorities of the Member States are responsible for implementing, supervising and enforcing it.

For a small business that mostly uses ordinary tools, two things are relevant. First, AI literacy (Article 4): organisations that use AI must make sure their people know enough about it. Second, transparency: the Commission writes that people must know they are talking to a machine when they use a chatbot, and that certain AI-generated content, such as deep fakes and text published to inform the public on matters of public interest, must be clearly labelled. The Commission also writes that the majority of AI systems currently used in the EU fall under minimal or no risk. The rules have recently been adjusted through a simplification package (the “AI Omnibus”), so check the current state on the Commission's page. This is not legal advice; if in doubt, consult a lawyer.

How to choose where to start

A sober order:

1. Choose one task. One that recurs, carries little risk and whose result you can easily check.
2. Decide what you enter. No personal data, confidential quotes or passwords unless you know what happens to them.
3. Keep a person ultimately responsible. You decide what goes out.
4. Measure before and after. How long did the task take before? Without your own measurement you do not know whether it helps.
5. Write down the rules. Put in a short note what the team does and does not do. That also helps with AI literacy.
6. Evaluate after a few months. Does it still fit? Then continue, otherwise stop.

What do we do?

We use tools to work faster, but every text and every piece of code goes through our hands before it goes live. We make no promises about time savings for your business without knowing your situation. If you want to know where AI or automation fits in your business, see our automation service or our coaching.

A thought example

This is also an invented example, not a customer case and not a result. Suppose a small consultancy writes the same kind of quote email every week. It decides to choose one task: a first draft of that email. Before it starts, it notes how long the task takes now. It enters no names or amounts, has a colleague review the draft and measures again after two months. If it turns out to make no difference, or if the review takes as long as writing it themselves, it stops. The point of the example is not the outcome but the order: one task, a measurement beforehand, a person who checks.

A short usage agreement

An agreement for your team need not be longer than half an A4 page. Think of: which tools are allowed, which data is never entered, who checks what goes out, and who to ask when in doubt. Such a note also helps when someone asks how you deal with AI, for example a customer who wants to know.

What do you do with AI on your website?

If you have a chatbot on your site, tell visitors they are talking to a machine; the European Commission names this as a transparency rule. If you publish AI-made texts, check them as you would check any text, and add your own experience. For imagery the same applies: use nothing whose rights you do not know. See also logo and brand basics.

Common mistakes

• Starting with the tool instead of the problem.
• Publishing unchecked. A language model can make things up.
• Entering personal data without knowing what happens to it.
• Wanting everything at once. One use at a time is manageable.
• Not naming an owner. Who is responsible when it goes wrong?

In conclusion

AI is a tool. The CBS figure shows that more and more companies use it, but not that every company needs it. Want to look together at what fits you? Book a free intro call; we are in no hurry.

Sources

Reviewed on 28 October 2026 against the sources below. Rules and guidelines change; this article is general information, not legal or tax advice.

• CBS: Bedrijven gebruiken AI het vaakst voor marketing of verkoop (2025)
• Europese Commissie: AI Act (regulatory framework for AI)
• Verordening (EU) 2024/1689 (AI Act), EUR-Lex
• Google Search: guidance on using generative AI content
• Verordening (EU) 2016/679 (AVG/GDPR), EUR-Lex
• Nationaal Cyber Security Centrum (NCSC)

Book an intro call