guides

Writing an AI acceptable-use policy your staff will read

Your staff are already using AI tools. The question is whether they know what not to paste into them. A long policy nobody reads does not answer that. One page does.

What the policy has to cover

Approved tools. Name the AI tools staff may use for work, and which accounts. A business account with data controls is not the same as a personal free account, even for the same product.

Data rules. Spell out what never goes into an AI tool: customer personal data, financial records, passwords and keys, anything under NDA, and health information. Give examples from your own business.

Checking the output. AI tools produce confident mistakes. Anything sent to a customer, filed with a regulator or used in a decision gets reviewed by a person who understands it.

Disclosure. Decide when staff must say that AI helped, for example in client deliverables.

New tools. Give staff one place to ask before trying a new AI tool, and a quick answer so they do not route around you.

Who owns it. Name the person who keeps the policy current and answers questions.

What to leave out

Skip the history of AI, long definitions and anything that will be out of date in three months. Name tools in an appendix you can update without redrafting the policy.

Rolling it out

  1. Send the one-page policy with a two-sentence note on why it exists.
  2. Run a 30-minute session with real examples from your business: good uses, bad uses, and the gray area.
  3. Configure the approved tools so the safe choice is the easy one.
  4. Review the policy every six months, and whenever a new tool becomes popular with your staff.

An AI acceptable-use policy is included in both of our retainers, and our AI implementation service covers the rollout, training and governance around it.

Need help with ai implementation?

Use-case selection, secure deployment, staff training, acceptable-use policy and governance.