5 Ways Insurance Agencies Can Reduce AI-Driven E&O Exposures

• September 9, 2026

By Timothy Ventura and Dana A. Gittleman

Insurance agencies and brokerages, like most businesses, are increasingly using artificial intelligence (AI) and large language models (LLMs), such as ChatGPT and Gemini, in their daily operations. While the advantages of AI can be significant— including greater efficiency, enhanced productivity and cost savings—these benefits come with risks.

AI-generated information may be inaccurate, leading to improper advice, coverage misrepresentations or binding coverage that violates underwriting guidelines. An overreliance on AI may also increase exposure to errors & omissions claims and litigation.

Insurance agencies can reduce AI-related risks by following these five best practices:

1) Require licensed producer oversight. AI can improve efficiency, but it does not replace the professional judgment or responsibility of a licensed producer. Treat AI-generated content as a preliminary draft that may contain errors.

Before any quote, proposal, coverage summary or policy comparison is provided to a client, a qualified individual should independently verify all material information, including coverage terms, limits, deductibles, exclusions and endorsements, against the applicable policy forms and source documents.

2) Establish a formal AI policy. Agencies should adopt a written AI policy that defines:

• Approved AI tools.
• Information that may or may not be entered into AI platforms.
• AI outputs requiring licensed producer review.
• Tasks AI may assist with or perform autonomously.
• Permitted and prohibited uses.

Approved uses may include document summarization and administrative support. Prohibited uses should include independently binding coverage, modifying limits, providing coverage advice or communicating coverage determinations without human review. Distribute the AI policy to all employees, support it with training and reinforce it through written acknowledgment of compliance.

3) Manage third-party AI and data security risks. Confidential client information, such as loss runs, financial records, proprietary information or other sensitive data, should never be entered into unapproved public AI platforms.

Agencies should also review contracts with AI vendors to ensure they adequately address confidentiality, cybersecurity controls, data retention, indemnification, insurance requirements and limitations of liability.

4) Document AI usage. As with other client communications and insurance placement decisions, agencies should maintain records of AI-related activity.

Documentation should reflect the client’s requests, information available to the producer, AI-generated output, any modifications or verification performed by agency personnel, communications with the client and the coverage ultimately procured. Your records may provide valuable evidence in defending a future E&O claim.

5) Review insurance coverage. Agencies should review their own E&O policies to determine whether any exclusions, limitations or endorsements affect coverage for AI-related activities. Likewise, agencies should be aware of any AI-related coverage restrictions contained in policies they recommend or place for clients.

Timothy G. Ventura is chair of the Insurance Agents and Brokers Liability Practice Group and the Professional Liability Practice Group in the Philadelphia office of Marshall Dennehey. Dana A. Gittleman is chair of the Real Estate E&O Liability Practice Group in the Philadelphia office of Marshall Dennehey.

This article was originally published on PLUS Blog, the blog of the Professional Liability Underwriting Society, on August 20, 2026. All rights reserved. Further duplication without prior permission is prohibited.