How to talk about AI with a compliance team

The short answer: bring compliance into the AI conversation before anyone picks a tool, not after something goes out the door. Start with what the team is already doing, name the uses you want by risk, ask what compliance needs to see, agree on a data rule, and set a date to review it together.

Why bring compliance in early?

When compliance first hears about AI from a flagged communication, the conversation starts defensive. When they help write the rules, they become the reason the program works. Early involvement turns a gatekeeper into a co-owner.

Compliance is not the brake on AI. It is the reason you can press the accelerator.

A five-part script for the first conversation

1. Start with what is already happening. “Here is how our team is using AI today, including the tools we have not approved.” Honesty sets the tone.

2. Name the uses you want, by risk. Separate internal work like drafting and summarizing from anything that reaches a client.

3. Ask what they need to see. “What would review and recordkeeping have to look like for you to be comfortable?”

4. Agree on the data rule. Which tools can touch client information, and which never can.

5. Set a review date. Agree to revisit the rules in a quarter, once you have real examples.

What to bring to the meeting

• A list of the AI tools in use and who uses them.

• Two or three specific use cases you want to approve first.

• A draft of how review and approval would work.

• Any vendor documentation on data handling and retention.

What to avoid

• Presenting a finished decision and asking for a signature.

• Leading with client-facing uses before internal ones are proven.

• Treating the policy as permanent. The tools will change, and so should the rules.

For the full framework, read compliant AI for financial advisors and our compliance-first guide to AI in financial services marketing.