Databrain

Simple rules for using AI at work

AI governance means agreeing who can use which tools, for what, with what information and how the results are checked.

Set clear boundaries

List approved tools and tasks. Explain what staff can share, when a person must review the output and which actions need permission. For example, a draft customer email might require approval before sending.

Review data and suppliers

Assess the information involved, the provider’s handling of it, retention, access and security. Do not put confidential or personal information into a tool unless the use has been assessed and authorised. Removing identifying details may not always make data anonymous.

Assign responsibility

Name an owner for each tool or process. Keep a list of tools, access and purposes, and agree how to report mistakes or suspected incidents.

Teach, test and revisit

Show staff how to use tools and check outputs. Review rules and supplier terms when the tools or their use change. These practices support responsible use; they do not by themselves prove legal compliance.

Make the rules usable

Keep a short register of approved tools, their owners, the work they support and the information they may use. For each important workflow, note where a person reviews the result, how access is removed and who handles a suspected mistake or incident.

Walk through a realistic example with the team, such as preparing a customer update from approved notes. Check whether the guidance explains what to do when information is missing or a tool gives an uncertain answer. Revisit it when the workflow or supplier changes.

The ICO’s AI and data protection guidance can help identify questions about personal information.

What would you like
more time for?