Cover image
Back to Blog

A One-Page AI Policy for a Small Team (Copy It, Then Change It)

•5 min read•AI Governance

Ask around your office and you'll probably find that someone already uses ChatGPT to tidy emails, someone else uses Gemini to summarise documents, and nobody has asked whether that's fine. Often it's on a personal account.

That isn't a reason to panic. It's a reason to write one page.

Below is a short AI policy a small team can copy, followed by why each line is there. Change the names and tools to fit your business. (It's a starting point, not legal advice.)

The policy

[Business name]: how we use AI at work

  1. Tools. For work we use [tool, e.g. ChatGPT Team / Claude Team / Microsoft Copilot] on our business account only. No personal accounts for work, and no new AI tools or browser add-ons without asking [name].
  2. Never put in: passwords, bank or card details, ID numbers, health information, or anything a client has asked us to keep private.
  3. Client information: only what the task needs, and only in the business account. Take out names and details when you can.
  4. AI drafts, a person decides. Anything that goes to a client, mentions money, or makes a promise is read and sent by a person.
  5. Check the facts. AI can make up figures, names, laws and quotes. Check anything you'll rely on.
  6. Say when it's AI. If a customer is talking to a chatbot or an AI voice, they're told it's AI.
  7. Mistakes happen; tell us fast. If something went in that shouldn't have, tell [name] the same day. No blame for reporting it.

Reviewed: [date]. Questions: [name].

That's it. Print it, send it round, go through it in a ten-minute team meeting.

Why each line is there

1. One tool, business accounts. This is the line that matters most. Netskope's cloud security data for October 2024 to October 2025 found that 47% of people using generative AI at work did so through personal accounts their companies weren't overseeing, as reported by Cybersecurity Dive. On a personal account, the business can't see what was shared, can't set data controls, and can't remove access when someone leaves.

2. The "never" list. Keep it short so people remember it. In a Cybernews survey of more than 1,000 US employees, 59% said they use AI tools their employer hasn't approved, and three-quarters of those admitted sharing possibly sensitive information with them, according to the Journal of Accountancy. Most of that isn't malice; it's people trying to get work done with no rule to follow.

3. Only what the task needs. "Rewrite this email to Mrs Papadaki about her overdue invoice" works just as well as "Rewrite this email to the client about the overdue invoice". Teach the habit of leaving the details out.

4. A person decides. AI is very good at drafts and very bad at knowing when a sentence will upset a client or commit you to something. The rule is simple: if it leaves the building, a person has read it.

5. Check the facts. AI tools sometimes invent things with complete confidence. The European Commission itself names this risk ("hallucination") as something staff using ChatGPT should be told about.

6. Say when it's AI. If you put a chatbot on your website or use an AI voice on the phone, people should know they're not talking to a person. In the EU this is becoming a legal duty for chatbots; elsewhere it's simply honest.

7. Tell us fast, no blame. If people are afraid of trouble, they hide mistakes, and the mistake you don't know about is the expensive one.

What the EU AI Act asks of a small business

The EU AI Act sounds like something for tech companies. One part of it isn't. Article 4, "AI literacy", applies to organisations that use AI systems, not only those that make them, and it has applied since 2 February 2025. National authorities can supervise and enforce it from 2 August 2026, according to the European Commission's AI literacy Q&A.

What the Commission's Q&A says, in plain words:

  • It covers everyday tools. Its own example is a company whose staff use ChatGPT to write adverts or translate text; they should be told about the specific risks, such as made-up answers.
  • No test, no certificate. There's no duty to measure how much staff know, and "there is no need for a certificate".
  • Keep a simple record. Organisations can keep an internal record of training or guidance. For a small team, the signed-off policy above, plus a note of the date you went through it together, is a sensible start.
  • Help exists. The Commission points small businesses to the European Digital Innovation Hubs as a first help desk on the AI Act.

We're not lawyers, and this isn't legal advice: if your business uses AI for anything high-stakes (hiring, credit, health), get proper legal help. But for most small firms, a clear page and a short conversation with the team is the honest first step.

A ten-minute rollout

  1. Pick the tool and set up the business account (or decide you'll do it this month).
  2. Fill in the brackets in the policy above.
  3. Go through it together, ten minutes, and ask everyone which AI tools they already use. You'll learn something.
  4. Write down the date and keep the page somewhere everyone can find it.
  5. Look at it again in six months. Tools change fast.

If you'd like help choosing the tool, setting up business accounts, and writing a policy that fits how your team actually works, that's what our AI Check-up is for: a 60-minute call and a one-page plan. If the team needs more than a page, the AI Team Workshop is a half day on your own tools and tasks.

Share:XLinkedInFacebook

Not sure where AI would save your business time?

A 60-minute call and a one-page plan naming the 3 jobs in your business worth automating first. Taken off the price of any build booked within 30 days.