- Your staff are almost certainly using AI already, often on personal accounts. A policy doesn't stop that. It makes it safe.
- Every policy needs the same 10 parts. The two that matter most: which tools are approved, and what must never be pasted into them.
- Keep it to one or two pages, have every person sign it, and review it every six months. The free template below is ready to adapt.
When we start work with a new client, one of the first things we ask is who on the team uses AI. The owner usually guesses two or three people. The real number, once we ask the staff, is almost always higher. People use free ChatGPT on their phones to write emails, summarize documents and answer customer questions. They don't mention it because nobody told them whether it was allowed.
That's the problem an AI policy solves. Not stopping people from using AI, which doesn't work, but telling them which tools to use, what's off limits and when to check with someone. This page explains what to include and gives you a template you can copy, edit and hand out this week.
Why you need one now
- Client data leaves your control. When someone pastes a client's details into a personal AI account, the business has no record of it and no way to delete it.
- Mistakes go out under your name. AI tools sometimes state wrong facts with total confidence. Without a review rule, those errors reach customers.
- You can't remove access. When an employee leaves, their personal AI account, and everything in it, leaves with them.
- Your best people hold back. Staff who aren't sure it's allowed either hide their use or avoid AI altogether. Both cost you.
- Clients and insurers are starting to ask. Larger customers increasingly include AI questions in vendor and security reviews. A written policy is the simplest answer.
The 10 sections every AI policy needs
| Section | What it answers |
|---|---|
| 1. Purpose and scope | Why the policy exists and who it covers (employees, contractors, interns) |
| 2. Approved tools | Which AI tools staff may use for work, on which accounts |
| 3. Never enter | The data that must never go into any AI tool |
| 4. Review before use | Who checks AI output before it reaches a client or the public |
| 5. Disclosure | When to tell clients that AI helped produce something |
| 6. Needs approval | Uses that require a manager's sign-off first |
| 7. Not allowed | Uses that are off limits entirely |
| 8. Accounts and access | Company accounts only, how access is granted and removed |
| 9. Mistakes and reporting | What to do if something sensitive was entered or wrong output went out |
| 10. Training and review | Required training, who owns the policy, when it's updated |
Free AI acceptable use policy template
Copy the text below into a document and replace everything in [brackets]. It's written for a business of 5 to 50 people using ChatGPT, Claude, Copilot or Gemini. Remove anything that doesn't apply.
[Company Name] AI Acceptable Use Policy
Effective date: [date]. Policy owner: [name, title]. Next review: [date six months out].
1. Purpose and scope
[Company Name] encourages staff to use AI tools to work faster and better. This policy explains how to do that safely. It applies to all employees, contractors and interns who use AI tools for company work, on any device.
2. Approved tools
You may use only the following AI tools for company work, and only through accounts provided by the company:
- [Tool, for example ChatGPT Business or Claude Team], accessed with your company login
- [AI features inside software we already use, for example Microsoft 365 Copilot or Gemini in Google Workspace]
- [Any other approved tool]
Personal or free AI accounts may not be used for company work. To request a new tool, ask [policy owner].
3. Never enter into any AI tool
- Passwords, login codes or security keys
- Social Security numbers, driver's license or passport numbers
- Bank account, credit card or payment details
- [Health information about clients or patients]
- Employee personnel, payroll or medical records
- Anything a client has told us is confidential, or anything covered by an NDA
- [Other sensitive categories specific to our business]
When you need help with a document that contains this information, remove or replace it first (for example, "Client A" instead of a name).
4. Review before use
AI can be wrong while sounding certain. You are responsible for anything you send, publish or act on, whether or not AI helped write it. Before AI-assisted work goes to a client or the public, check every fact, number, name, date and quote. [Contracts, legal language, financial figures and pricing must also be reviewed by name or role.]
5. Disclosure
Using AI to draft, edit or summarize is normal and doesn't need to be disclosed. Tell the client when [AI produced a substantial part of a deliverable they are paying for / a client contract requires it / AI was used to analyze their data]. Never present AI output as a person's personal review if no person reviewed it.
6. Needs approval first
- Connecting an AI tool to company email, files, CRM or other systems
- Uploading client documents or data sets
- Using AI to communicate with customers automatically (for example, chatbots or automatic replies)
- Creating images, video or audio of real people
Ask [policy owner or manager] before doing any of the above.
7. Not allowed
- Using AI to make hiring, firing, pay or discipline decisions
- Using AI to give professional advice [legal, medical, tax, investment] without review by a qualified person
- Creating content that impersonates a real person or company
- Recording or transcribing meetings or calls without the knowledge of the people on them [and as required by law]
- Anything that breaks the law or another company policy
8. Accounts and access
[Policy owner] sets up AI accounts and removes them when someone leaves or changes role. Don't share logins. Company AI accounts and everything in them belong to [Company Name].
9. Mistakes and reporting
If you entered information covered in section 3, or AI-assisted work with an error went to a client, tell [policy owner] the same day. Reporting quickly is what matters. Nobody will be penalized for an honest, promptly reported mistake.
10. Training and review
Everyone completes [company AI training] within [30 days] of starting or of this policy taking effect. [Policy owner] reviews this policy every six months and when we add a new tool.
Acknowledgment
I have read and understand this policy. Name: ________ Signature: ________ Date: ________
How to roll it out so people follow it
Ask before you write
Run a no-blame, two-minute survey: which AI tools do you use, for what, on which account? You'll learn what to allow, not only what to ban.
Give people a good tool first
A policy that only says "no" pushes use underground. Pay for a business plan and approve it in the same week you publish the policy. Our comparison of the main business AI tools will help you choose.
Walk the team through it
Fifteen minutes in a team meeting, with real examples from your business: "This client email is fine. This one has an account number, so remove it first."
Get signatures
Add it to onboarding for new hires and keep the signed copies with other HR records.
Review every six months
Tools and features change fast. Put the review date on the calendar now.
Extra rules for regulated industries
The template covers the basics. Some industries need more. This isn't legal advice, so have your attorney or compliance officer review your final version.
Healthcare, dental and med spas
Protected health information shouldn't go into any AI tool unless the vendor has signed a Business Associate Agreement (BAA) with your practice under HIPAA, and the tool is set up the way that agreement requires. Coverage differs by vendor and plan. As of November 2026, OpenAI doesn't offer a BAA for ChatGPT Business, and Anthropic's BAA doesn't cover Claude Team. Both offer one only on enterprise or healthcare tiers. Microsoft's BAA can cover Copilot used with work accounts, and Google's can cover Gemini in Workspace once an admin accepts it. Confirm coverage with the vendor in writing. Until it's in place, staff should remove patient identifiers before using AI. Our guide to compliance-first AI for dental practices goes further.
Financial advisors and insurance agencies
Business communications with clients are often subject to recordkeeping and supervision rules, and that can include AI-drafted messages. Client account details belong in the "never enter" list. Check with your compliance team or broker-dealer about approved tools. See compliance-first AI for financial advisors.
Law firms
In January 2024 the Florida Bar issued Ethics Opinion 24-1 on generative AI. It says lawyers may use it but must protect client confidentiality, supervise the output as they would a nonlawyer assistant, and bill fairly for AI-assisted work. It also recommends getting a client's informed consent before sharing their confidential information with a third-party AI tool. Other state bars have issued similar guidance. Your policy should reflect your bar's opinion, and review rules should name an attorney.
Common mistakes we see
- Banning AI outright. It doesn't stop use. It just stops you seeing it.
- A ten-page policy. Nobody reads it, so nobody follows it.
- No approved tool. Rules without a sanctioned option send people back to personal accounts.
- Writing it once. A policy from 2024 probably doesn't mention the AI features now built into your email and documents.
- No owner. Someone has to approve tools, remove access and answer questions. Name them.
Common questions
Does a small business need an AI policy?
Yes, if anyone on your team uses AI tools for work, and most already do. A one or two page policy that names the approved tools and lists what must never be pasted into them removes most of the risk. It also gives staff permission to use AI properly instead of quietly.
What should an AI acceptable use policy include?
Purpose and scope, approved tools, data that must never be entered, rules for reviewing output, disclosure to clients, uses that need approval, account and access rules, how to report a mistake, training, and a review date. The template on this page covers all ten.
Can employees use ChatGPT at work?
They can if the business allows it and sets rules. We recommend a company business plan rather than personal accounts, so the business controls the data, can see who has access and can remove it when someone leaves.
How long should an AI policy be?
For most businesses under 50 people, one to two pages. A long policy nobody reads protects you less than a short one everyone signs and understands.
Is this template legal advice?
No. It's a practical starting point written by an AI consulting firm. Businesses in regulated fields such as healthcare, financial services and law should have their attorney or compliance officer review the final version.
This article is general information, not legal, tax or compliance advice. Prices, plans and vendor terms change; confirm current details with each vendor before you rely on them. ChatGPT, Claude, Microsoft 365 Copilot, Gemini and other product names are trademarks of their owners. FlowBack AI is not affiliated with or endorsed by these companies except as stated.
Carter Simmons is the founder of FlowBack AI, an AI consulting firm in Tampa. He holds 13 AI certifications from Anthropic (the maker of Claude), Google and Microsoft, and FlowBack AI is an OpenAI Select Partner. The firm runs AI Opportunity Audits, advisory programs, implementation and team workshops for established businesses in Tampa Bay and nationwide.
Newsletter
The Owner’s AI Brief
Twice a month for owners of established businesses: one AI tool you can use in 15 minutes, with the dollar math behind it.
Subscribe to the newsletterWant the policy and the training done for you?
We write AI policies with clients as part of our audits and workshops, then train the team on them in person. Book a free 15-minute call to see whether it fits.
Speak with a consultant →