2026-06-04 · 8 min read
The AI Acceptable Use Policy Your Company Actually Needs (Free Template)
Most companies discover they need an AI acceptable use policy the same way: someone pastes a customer list into a free chatbot, someone else publishes AI-written copy with an invented statistic, and suddenly “we should probably have rules for this” becomes urgent. The good news: a policy that prevents 90% of the damage fits on one page. This post gives you that page — copy it, adapt it, ship it this week.
Why a policy beats a ban
Blanket bans don't stop AI use; they just push it underground. Industry surveys consistently find that a large share of employees use AI tools their employer never approved — so-called shadow AI — and unsanctioned use is invisible use: no data rules, no quality control, no idea what went into which output. A short, permissive-but-clear policy gets usage into the open where you can manage it. (We make the longer argument in AI transformation is a governance problem.)
The one-page AI acceptable use policy (template)
Adapt the bracketed parts. Written for a 5–50 person company; delete what doesn't apply.
[COMPANY] AI Acceptable Use Policy — v1.0, [DATE]
1. Approved tools. Employees may use the tools on our approved list [link to internal list] for work tasks. Want a new tool approved? Ask [OWNER] — answer within [3] working days. Tools not on the list may be used only with data that is already public.
2. Data rules (the important part).Never enter into ANY AI tool: customer personal data (names, emails, account details), employee data, financial records, credentials or API keys, unreleased product information, or anything covered by an NDA. When in doubt: if you wouldn't paste it into a public forum, don't paste it into an AI tool. Tools with an enterprise/no-training agreement [list them] are exempt for the data classes named in that agreement.
3. Output rules. You own what you ship. AI output must be reviewed by a human before it reaches a customer, a public channel or a business decision. Verify every fact, number and citation — AI tools state falsehoods confidently. AI-generated content that goes out under our name follows the same quality bar as human-written work.
4. Disclosure. We disclose AI use where it matters: [e.g. AI-assisted content is reviewed by a named editor; customer-facing chatbots identify themselves as automated; we never present AI output as independent human opinion in reviews or testimonials].
5. Accounts. Use work accounts (work email) for work AI use — no personal accounts. This keeps access manageable when people join or leave.
6. Ownership & questions. [OWNER] owns this policy, maintains the approved-tool list, and reviews this document every [quarter]. Mistakes happen — report a data slip to [OWNER] immediately; fast reporting is never punished.
How to roll it out without the eye-rolls
Name one owner, not a committee.For a small company that's usually whoever runs ops or IT. Their job is the approved list and a fast yes/no on new tools — slow approvals recreate shadow AI.
Start the approved list with what people already use. Survey first (amnesty included), then approve the sensible tools the team has adopted. Our free-tier reviewsflag which vendors train on your data on free plans — that's frequently the deciding factor between a free tool being fine or a data risk.
Make the data rule memorable, not exhaustive.“Don't paste anything you wouldn't post publicly” survives in someone's head at 5pm on a Friday. A 14-category data classification matrix does not.
Revisit quarterly.Tools change their data terms (we track free-tier changes monthly for exactly this reason). A policy dated 18 months ago tells your team the policy doesn't matter.
What this policy deliberately doesn't do
It doesn't try to be EU AI Act compliance, a security review process, or a model risk framework — at 5–50 people you need behavioral guardrails, not a governance department. If you handle regulated data (health, finance, children), treat this as the floor and get sector-specific advice on top.
Choosing tools after the policy?
Every tool we recommend comes with the exact truth about its free tier — start with the free AI marketing tools that are actually free.