Write an AI Policy for Your Team or Brokerage (Template Included)

If your team uses AI and you don't have a written policy, you still have a policy. It is whatever each person decided on their own. One agent pastes a client's financials into a free chatbot, another publishes an AI-written ad without reading it, and nobody finds out until something goes wrong.

A short written policy fixes that. It doesn't need to be a legal document. It needs to be clear, brief, and actually used.

If you haven't read it, my AI governance guide covers the thinking behind this one.

Why Now

• State laws are accumulating. Rules on chatbots, deepfakes, and altered listing photos are being passed state by state. See our state AI laws tracker.
• Liability sits with you. "The tool did it" rarely helps. We looked at this in Who Is Liable When an AI Agent Makes the Mistake?.

What a Good Policy Covers

Keep it to one or two pages with these sections.

1. Approved Tools

List the AI tools your team may use for work, and on which plans. Say which ones are not approved. Check with your broker or compliance contact for anything that touches client data.

2. What Never Goes In

Define what can't be entered into any AI tool unless that tool is specifically approved for it:

• Social Security numbers, bank account or wire details
• Full financial statements, tax returns, or credit information
• Government ID numbers
• Anything a client asked you to keep confidential

When in doubt, replace names and addresses with placeholders.

3. Human Review

Everything AI produces gets read by a person before it goes out, and that person is responsible for it. Legal or contract language goes to the broker or an attorney.

4. Marketing and Fair Housing

Require a fair housing check of listing copy and ads. See my fair housing and AI checklist. State your rules on altered photos and virtual staging, including labeling.

5. Client Communication and Disclosure

Say when and how you tell clients that AI is being used, such as with chatbots, texting assistants, or call-recording tools. Automated messages should identify themselves where required.

6. Fraud Safeguards

AI makes impersonation easier. Require a phone call to a number you already have before acting on any change to wire instructions. See my guide on AI wire fraud and deepfakes.

7. Record Keeping

Decide what to save, such as final published versions and client-facing AI summaries. Follow your state's retention requirements.

8. Incidents

If someone enters restricted information in an unapproved tool or publishes something incorrect, they report it right away to a named person.

9. Review Schedule

Review the policy every six months. The tools and the rules change quickly.

A Starter Template

Copy this and fill in the brackets:

[Team / Brokerage Name] AI Use Policy
Effective: [date]   Owner: [name]   Next review: [date]

1. APPROVED TOOLS: [list tools and plan types]. No other AI tools may be used with client information.
2. NEVER ENTER: SSNs, bank/wire details, full financial documents, government ID numbers, or confidential client information. Anonymize when possible.
3. HUMAN REVIEW: All AI output is reviewed by the person sending it before use. That person is responsible for its accuracy.
4. MARKETING: All listing copy and ads follow fair housing rules. Altered or staged photos are labeled, and originals are kept.
5. CLIENT DISCLOSURE: We tell clients when automated tools communicate with them. [Describe your wording.]
6. FRAUD: Wire or payment changes are verified by phone using a known number before any action.
7. RECORDS: We keep final versions of published content for [period].
8. INCIDENTS: Report problems to [name] immediately.
9. REVIEW: This policy is reviewed every six months.

Making It Stick

• Keep it short. Nobody follows a 12-page document.
• Walk the team through it using real examples.
• Have everyone acknowledge it.
• Update it when something changes, like a new tool or a new state law.

Of course, this template is a starting point, not legal advice.

- Jason