A one-page AI usage policy your team will actually read
When a team has no AI policy, people either avoid the tools or use them quietly with no rules. Both are bad. A policy doesn’t have to be a legal document. One clear page, written in plain words, does more good than twenty pages nobody opens.
Seven sections
Figure 1 · A one-page outline
| Section | What to write | Example line |
|---|---|---|
| Purpose | Why the policy exists | “Use AI to do better work, safely.” |
| Approved tools | Named tools and accounts | “Use the company account, not a personal one.” |
| Never share | Data that stays out of any tool | “No customer personal data or credentials.” |
| Check before use | What must be reviewed | “A person checks anything sent outside the company.” |
| Say when | When to disclose AI use | “Tell the client if AI drafted a deliverable.” |
| Ask | Who answers questions | “Questions go to the named owner.” |
| Review date | When it will be updated | “Reviewed every quarter.” |
What makes it work
- Be specific. “Use AI responsibly” tells people nothing. “Don’t paste customer data into any tool except the approved one” does.
- Name an owner. Someone has to answer questions and update the page.
- Explain the reasons. People follow a rule better when they know what it protects.
- Revisit it. Tools change quickly. A policy that is a year old is probably wrong.
- Pair it with practice. Rules land better when people have also practised the tools on real tasks.
Link it to the rest
Your policy should point to your privacy obligations, which we outlined for India in the DPDP note, and to the review points for agents in our note on human review. Part of our corporate training is helping a team agree its own version.