AI Use Policy Template for Small Business: A One-Page Version You Can Copy
By Jordan SolenderUpdated 9 min read
An AI use policy is a short written set of rules for how your staff may use AI tools at work: which tools are approved, what information may go into them, who checks the output and who is responsible when something goes wrong. A small business does not need a long document. One page that people will read does more than twenty pages that nobody opens.
This guide gives you a one-page template you can copy, explains each section and shows how to roll it out. It is a practical starting point, not legal advice. If you handle regulated data, have your attorney review the final version.
Why a small business needs an AI use policy
Your staff are very likely using AI already, whether or not you have approved it. Someone pastes a customer email into a chat tool to draft a reply. Someone else uploads a spreadsheet to get a summary. Most of that use is harmless and helpful. The problem is that nobody has said where the line is, so each person draws their own.
A written policy fixes four things:
- Data exposure. People know what must never be pasted into an AI tool.
- Tool sprawl. The company uses a short list of approved tools on business accounts instead of a dozen personal ones.
- Unchecked output. Anything AI writes for a customer is read by a person first.
- Ownership. One named person answers questions and approves new tools.
It also helps with sales. Larger customers and cyber insurance applications increasingly ask whether you have rules for AI use. A one-page policy lets you answer yes and show it.
What an AI use policy should cover
A complete policy for a small business has seven parts. Each one answers a question your staff would otherwise have to guess at.
| Section | The question it answers |
|---|---|
| Purpose and scope | Who does this apply to, and to which tools? |
| Approved tools | What may I use, and on which account? |
| Data rules | What may I put into an AI tool, and what may I never put in? |
| Human review | When must a person check the output? |
| Disclosure | When do we tell customers that AI was involved? |
| Ownership | Who do I ask, and who approves a new tool? |
| Incidents and review | What do I do if something goes wrong, and when is this updated? |
The one-page AI use policy template
Copy the text below and replace everything in brackets. Keep it to one page. If you find yourself adding paragraphs, move the detail to a separate procedure and keep the policy short.
[Company name] AI Use Policy Effective [date]. Owner: [name and role]. Reviewed every [six] months.
1. Purpose and scope. We encourage the use of AI tools to do better work, faster. This policy applies to all employees and contractors, and to any AI tool used for company work, including chat assistants, AI features inside software we already use, meeting recorders and AI agents.
2. Approved tools. Use only the tools on the approved list, signed in with your company account: [list tools]. Do not use personal accounts or free versions of AI tools for company work. To request a new tool, contact the policy owner before you use it.
3. Data rules. You may use approved tools with ordinary business information. Never enter the following into any AI tool unless the policy owner has approved that tool for that data in writing:
- Passwords, API keys or other credentials
- Payment card numbers, bank account numbers or Social Security numbers
- Health information about any person
- Customer information covered by a confidentiality agreement
- Employee records, including pay, reviews and medical details
- [Other categories specific to your business]
4. Human review. You are responsible for work you produce with AI, the same as any other work. A person must read and approve AI output before it is sent to a customer, published, used in a contract, proposal or invoice, or used to make a decision about hiring, pay or discipline. Check facts, figures and quotations against a reliable source. AI tools can state incorrect things with confidence.
5. Disclosure. We do not present AI-generated content as the personal work of a named person without their review. When a customer talks to an automated AI agent by phone or chat, the agent identifies itself as an AI assistant. Tell meeting participants before you record or transcribe a meeting with an AI tool.
6. Ownership. [Name] owns this policy, keeps the approved tool list, answers questions and approves exceptions. Managers are responsible for making sure their teams follow it.
7. Incidents and review. If you believe restricted information was entered into an AI tool, or AI output caused an error that reached a customer, tell the policy owner the same day. Reporting quickly is expected and will not be held against you. This policy is reviewed every [six] months and whenever we adopt a new AI system.
Acknowledged by: [employee name, signature, date]
How to fill in each section
Approved tools
List tools by name, and say which plan. The plan matters because business and enterprise plans of the major AI tools usually come with different data handling terms than free consumer versions. Read the vendor’s current terms for the plan you pay for, and check two things: whether your inputs are used to train models, and how long they are kept.
Do not forget AI features inside software you already own. Your CRM, help desk, email client and meeting software may all have AI turned on. Include them on the list or turn them off.
Data rules
This is the section that matters most. Write it in terms of things people recognize, not legal categories. “Health information about any person” is clearer than a citation to a statute.
Then add what is specific to your industry:
- Healthcare and related businesses: patient information generally may not go into a tool unless the vendor has signed a business associate agreement with you.
- Law, accounting and financial firms: client files and anything privileged or covered by professional confidentiality rules.
- Government contractors: controlled or contract-restricted information, which usually has its own handling rules.
- Any business: information a customer contract says you must keep confidential. Check your larger customer agreements, because some now restrict AI use directly.
State privacy laws may also apply to personal data you hold about consumers. Your attorney can tell you which ones cover your business.
Human review
Decide where a person must be in the loop and name those points. Customer-facing messages, anything with a price or a legal commitment, and decisions about people are the usual three. Internal drafts, brainstorming and summaries for your own use normally do not need a second reader.
If you run AI agents that answer customers directly, review works differently. A person cannot read every reply before it goes out, so the control moves to the design: the agent answers only from approved sources, hands off to a person when unsure, and a sample of conversations is reviewed each week. Our guide on how to build an AI support agent covers those controls.
Ownership
Name one person, not a committee. In a company of 20, that is often the owner or the operations lead. They do not need to be technical. They need the authority to say yes or no to a tool and the time to answer questions within a day. If nobody has that time, a fractional Chief AI Officer typically writes and maintains this policy as part of the role.
Three levels of data, if you want one more layer
If the list in section 3 feels too blunt, sort your information into three levels. This stays simple enough to remember.
| Level | Examples | Rule |
|---|---|---|
| Open | Public website copy, published pricing, general research | Any approved tool |
| Internal | Internal emails, process documents, sales notes without sensitive details | Approved tools on company accounts only |
| Restricted | Credentials, payment and health data, employee records, confidential customer files | No AI tool unless approved in writing for that data |
Most daily work falls in the first two levels, which is the point. The policy should make ordinary use easy and make the restricted cases obvious.
How to roll it out
A policy that arrives as an email attachment gets ignored. Use these five steps.
- Ask what people already use. A five-minute survey with no penalty for honest answers. You cannot write an approved list without knowing the real one.
- Pick the approved tools and pay for business accounts. If you ban personal tools without giving people an approved alternative, they will keep using the personal ones quietly.
- Fill in the template. One page, the owner’s name on it.
- Walk through it in a meeting. Fifteen minutes, with two or three real examples from your own business of what is fine and what is not.
- Collect acknowledgments and set the review date. Put the next review on the calendar the same day.
Common mistakes
- Banning AI outright. People use it anyway, on personal accounts, where you have no visibility and no say over the data terms.
- Copying an enterprise policy. A document that refers to review boards and committees you do not have teaches staff that the policy is not meant for them.
- Writing it once. Tools and their terms change. A policy with no review date is out of date within a year.
- Leaving out AI inside existing software. The meeting recorder and the CRM assistant handle more sensitive information than most chat windows.
- No named owner. If nobody can approve a new tool, the answer to every request becomes silence, and people stop asking.
Where the policy fits in your AI plans
A policy is the second step, not the first. The first is knowing where AI would help. Our AI readiness assessment for small business gives you nine questions for that, and the free Automation Scorecard scores them for you in about three minutes. Once you know the first workflow you want to automate, the policy tells everyone the rules it will run under. The small business AI adoption playbook covers the steps after that.
Frequently asked questions
What is an AI use policy?
An AI use policy is a written set of rules for how employees may use AI tools at work. It names the approved tools, the information that may not be entered into them, when a person must review AI output and who is responsible for the policy. It is sometimes called an AI acceptable use policy or an AI usage policy.
Does a small business really need an AI policy?
Yes, if anyone on staff uses AI tools for work, and in most businesses someone does. The policy can be one page. Its job is to stop sensitive information from going into the wrong tool and to make sure a person checks anything AI produces before a customer sees it.
How long should an AI use policy be?
One page is enough for most small businesses. Seven short sections cover scope, approved tools, data rules, human review, disclosure, ownership and incidents. Longer documents are read less. Put detailed procedures somewhere else and link to them.
Who should write the AI use policy?
The person who will own it: usually the owner, the operations lead or whoever is responsible for IT. Involve one or two staff who use AI heavily, since they know the real use cases. If you handle regulated data, have an attorney review it before it takes effect.
Should we ban ChatGPT and similar tools at work?
A ban rarely works. Staff tend to keep using the tools on personal accounts, which gives you less control, not more. Approving a short list of tools on business accounts and setting clear data rules is usually the safer choice.
How often should we update the policy?
Review it every six months and whenever you adopt a new AI system, add an AI feature to existing software or take on a customer whose contract restricts AI use. Record the review date on the policy itself.
Is this template legal advice?
No. It is a practical starting point written for general business use. Laws on privacy, employment and specific industries vary by state and by sector. Have your attorney review your final policy, especially if you handle health, financial or government data.
If you would like help writing the policy or choosing the first workflow it will cover, book a strategy call. AI consulting for small business explains how we work with smaller companies.