The warning comes on the heels of a volatile period for insurance broker and agent stocks. On Feb. 9, the subsector plunged 9% following news that two digital insurance companies—U.S. auto comparative rater Insurify and Spanish homeowners insurer Tuio—had launched chatbot assistants utilizing ChatGPT technology. However, over the next three weeks, insurance distribution stocks rallied 7%, outpacing a broader S&P 500 decline of 1%. The marketplace appeared to digest the AI threat and decided it was not a material risk to revenue growth, adopting a broadly optimistic “nothing to fear” and “far away” sentiment.

BofA disagrees.

“Our view is that large language model digital agents can effectively do a non-immaterial portion of the work currently provided by 20-30k independent agents across the United States,” the BofA report stated.

The core of the firm’s bearish thesis centers on a massive pool of routine, low-complexity insurance policies. The BofA analysts, Joshua Shanker, Joseph Tumillo, Cyril Onyango, and Fatima Keita, looked at just six major carriers catering to small businesses and personal lines: Travelers, Hartford, Progressive, Cincinnati Financial, Hanover, and Selective. From these six companies alone, BofA identified over $15 billion in commissions paid to independent agents in 2025 that largely skew toward low-complexity risks.

For example, Progressive paid over $6 billion to independent agents last year, while Travelers and Hartford paid roughly $3.35 billion and $1.25 billion, respectively, in segments dominated by personal lines and small commercial business. BofA notes that these types of policies, such as standard home and auto insurance, represent low-sophistication transactions where human agents add little value, making direct-to-consumer digital channels a considerable cost-saver for the buyer.

Amrish Singh, CEO of the AI insurance startup Liberate, told Fortune that he thinks BofA’s estimate checks out. His own math shows a wide range of $4.8 billion to $33.6 billion of insurance tasks that can be automated in the U.S. alone.

The snowball effect

While bulls argue that large insurance brokers do not heavily participate in personal lines or small commercial markets, BofA counters that years of constant “tuck-in M&A” have created a “snowball effect.” Hundreds of small acquired shops have brought a significant amount of low-complexity, small-ticket business under the umbrellas of large brokers, a vulnerability that is often obscured by subpar public disclosures. Furthermore, even large-case, complex business—which is unlikely to face direct disintermediation—could experience pricing deflation as AI demystifies the insurance markets for sophisticated corporate buyers.

Some investors have equated the AI threat to the much-hyped but slow-to-materialize disruption of self-driving cars. However, BofA draws a sharp distinction. While transitioning to autonomous vehicles will require trillions of dollars in infrastructure and take many years, deploying large language model chatbots is cheap, easy, and happening right now. As an example, the report points to Munich Re’s Next Insurance, which already offers an AI chatbot on its site where customers can purchase and bind commercial policies directly without a human agent.

While acknowledging that making long-term predictions in the face of technological innovation is “difficult,” BofA notes that Facebook/Meta and Google/Alphabet did not replace print advertising overnight but that over 20 years, consumer behavior changed to dramatically shrink the print ads market. “We are not arguing that insurance intermediaries will disappear or that Coca-Cola will buy its insurance from a chatbot,” BofA said, but it urged investors to look closely at this sector, as insurance distributor stocks do not seem to be discounting the risks.

BofA points out that the sector currently trades at 22x trailing free cash flow and 15 times enterprise value to trailing Ebitda. While bulls might argue that the stocks look cheap after falling 24% from peak valuations set a year ago, BofA cautions that these multiples have merely returned to pre-pandemic levels. Furthermore, BofA asserts that insurance distribution firms frequently utilize liberal earnings “adjustments”—such as excluding integration costs from their steady stream of acquisitions—that tend to significantly flatter their true earnings power.

Ultimately, BofA is not predicting the overnight disappearance of the human insurance agent, nor is it suggesting that massive corporations like Coca-Cola will suddenly buy complex insurance policies from a chatbot. However, BofA warns that an agency business currently perceived as having 3% to 7% organic revenue growth could see that slip to 1% to 5% in the face of disruptive technology. BofA concludes that with 10% to 20% of current business potentially facing disintermediation, the industry’s premium valuations leave very little room for error.

For this story, Fortune journalists used generative AI as a research tool. An editor verified the accuracy of the information before publishing.