An AI policy for a small business fits on one page. It sets which tool your team uses, which data may go in, who checks results and when AI use is disclosed to customers. More important than any detail is that everyone knows it and the approved tool is really there.
Which tool is allowed?
Name one tool everyone works with, with a company account and a contract. A ban on private accounts only works if there is a good alternative.
Write down who grants access and who revokes it when someone leaves.
Which data may go into the chat?
Sort data into levels: public information may go in, internal information such as prices and procedures only through the company account, personal data only when the task requires it.
Health data and other special categories under Art. 9 GDPR belong in an AI chat only with a clear legal basis. When in doubt, they stay out.
Source: European Union, EUR-Lex: Regulation (EU) 2016/679, General Data Protection Regulation, 27 Apr 2016
Who checks what the AI writes?
A person, before anything leaves the house. AI answers can be wrong: according to Bitkom, 42 percent of those who use AI for search have already received false or invented information. Only 57 percent check the results before using them.
Decide which texts are always checked, for example everything to customers and everything with figures or legal effect.
Source: Bitkom: Internet search in transition: half already use AI chats, 20 Nov 2025
Do AI texts have to be labelled?
For certain cases the European AI Act requires transparency, for example when people interact with an AI system or when certain AI-generated content is published. Which of it concerns your business is a question for the individual case.
Regardless of that, a clear line builds trust. In all Agentiq products every AI message is labelled.
Source: European Union, EUR-Lex: Regulation (EU) 2024/1689, Artificial Intelligence Act, 13 Jun 2024
Who is responsible?
Name one person who answers questions, checks new tools and keeps the policy current. In a small business that is often the owner.
Go through the policy together with the team instead of just sending it round.
How does Agentiq Chat help?
One AI chat for the whole company, processed in the EU.
- All major models, processed in the EU
- Projects with your own knowledge and documents
- No training on your data
Price on request.
Frequently asked questions
Do very small teams need an AI policy too?
Yes, a short one. As soon as anyone in the team uses AI for work, a shared line on tool, data and checks helps.
Can I simply ban AI use?
You can. But many bring their AI tools anyway: according to Microsoft and LinkedIn, 78 percent of AI users at work do. An approved tool with clear rules gives this use a frame.
Is an AI policy required by law?
The law does not prescribe a particular form. The duties under the GDPR and the AI Act apply regardless. This text is not legal advice.
How does Agentiq Chat help?
Agentiq Chat gives your team a tool with a company account: all major models processed in data centres in the EU, projects with company knowledge and administration for admins. Your data is not used for training.
What does Agentiq Chat cost?
Agentiq Chat comes for up to 5, 25 or 100 people. We give the prices for Small, Medium and Premium on request. The purchase page always shows the current offer.
Sources
- European Union, EUR-Lex: Regulation (EU) 2016/679, General Data Protection Regulation27 Apr 2016
- Bitkom: Internet search in transition: half already use AI chats20 Nov 2025
- European Union, EUR-Lex: Regulation (EU) 2024/1689, Artificial Intelligence Act13 Jun 2024
- Microsoft and LinkedIn: 2024 Work Trend Index Annual Report8 May 2024
- Agentiq: Agentiq Chat: services and tiersas of 13 Oct 2026