How can an independent insurance agent use AI safely?

An independent insurance agent can safely use AI to draft emails, social posts, newsletters and meeting notes, as long as a licensed person reads and corrects everything before it goes out. Anything that promotes a policy, an insurer or your agency is treated as advertising under state rules no matter who or what wrote it, so the same advertising rules apply to an AI draft as to anything you write by hand.

What can AI safely do for an independent insurance agent?

AI is good at the writing around the sale. It is not a quoting, rating or coverage tool, and it can sound confident while being wrong.

Tasks that work well with a quick human check:

• Client emails. Renewal reminders, a thank you after a new policy, or a note asking for updated driver or property details.
• Social posts and newsletters. Seasonal tips, such as checking smoke detectors or reviewing a home inventory, written in general terms without naming a specific policy.
• Meeting notes. Turning your rough notes from a client meeting into a clean summary and a short list of follow-up tasks.
• Plain language explanations. A first draft that explains a term like deductible or replacement cost, which you then check against the actual policy form before anyone relies on it.
• Website pages. A first draft of your About page or a page explaining how an independent agency works.

Start with one or two of these and see whether they actually save you time.

Is an AI-written post or email advertising?

Often, yes. The NAIC model regulation on life insurance and annuity advertising (Model 570) defines an advertisement as material designed to create public interest in life insurance or annuities, an insurer or an insurance producer, and it lists the Internet, web pages and form letters. It also says every advertisement, "regardless of by whom written, created, designed or presented," is the responsibility of the insurer and of the producer who created or presented it. Under the same model, insurers must tell their agents at least once a year about their procedures for approving ads the insurer did not supply, so check your carrier agreements before you post.

The NAIC Unfair Trade Practices Act (Model 880) lists as unfair practices putting untrue, deceptive or misleading statements about insurance before the public, including by email or internet posting, and misrepresenting the benefits or terms of a policy.

These are model laws. Each state adopts its own version, so your state's wording is what counts. Texas is one example. The Texas Department of Insurance's adopted text of 28 TAC Chapter 21 defines an advertisement to include electronic media, form letters and material used to sell more coverage to existing clients. It also requires an ad to identify the person responsible for it. For an agent's ad that covers insurance in general rather than a specific policy, it is enough to state the agent's full licensed name, registered assumed name or Texas license number.

What should I never paste into a chatbot?

Keep these out of general AI tools unless your agency has approved the tool for client data:

• Names, addresses, dates of birth and Social Security numbers
• Driver's license numbers, VINs and claim numbers
• Health, financial or bank account details
• Declarations pages, applications and loss runs
• Carrier documents marked confidential

Replace names with placeholders like "Client A," and keep requests general. Also check the tool's data settings. OpenAI's ChatGPT data controls page says that on personal Free, Plus and Pro accounts, sharing chats to improve its models is on by default and can be turned off in Settings. It also says content in ChatGPT Business and Enterprise workspaces is not used for training by default. Other tools have their own settings, so read them before you start.

Does the NAIC AI bulletin apply to agents?

Not directly. The NAIC Model Bulletin on the use of AI systems by insurers, adopted December 4, 2023, is addressed to "all insurers licensed to do business" in a state. It asks insurers to have a written program for governing the AI systems they use in decisions that affect consumers, such as underwriting and claims. It does not set rules for an agent writing a newsletter with a chatbot.

That does not leave agents without rules. Your marketing still falls under your state's advertising and unfair trade practice rules described above, and carriers may set their own policies for appointed agents.

What should I check before an AI draft goes out?

Run every draft through a short review before you send or post it:

1. Facts. Every coverage statement matches the actual policy form. Remove anything you cannot confirm.
2. No promises. No guaranteed savings, approval or claim outcomes.
3. Identification. Your name or agency name, and anything else your state requires, such as a license number.
4. Insurer names. If you name an insurer or a specific policy, check your carrier's ad approval rules first.
5. Privacy. No client details that should not be shared.
6. Your voice. Read it out loud and rewrite anything you would not say to a client.
7. A record. Save a copy of what you posted and when, in case your carrier or state asks.

How do I find my own state's rules?

Start with your state department of insurance. The NAIC keeps a directory of state insurance departments with links to each one. Look for the advertising rules for the lines you sell, and ask the department or your carriers by email if something is unclear. Rules differ by state and by line of insurance, so do not assume the Texas example applies to you.

AI can take some of the typing off your plate, but the license on the ad is still yours.

If you want help with the marketing side of your agency, you can learn how Momyntum works or email hello@momyntum.com.

Updated October 31, 2026. For: Insurance.