Investors who dismissed last month’s artificial intelligence (AI) developments may need to prepare for a sobering reassessment. A new BofA Global Research report calculates that over $15 billion in insurance industry commissions are deemed “low complexity” and face a significant risk of being bypassed by AI. Put simply: it’s a genuine threat.
This caution follows a turbulent spell for insurance broker and agent stocks. On February 9, that subsector dropped 9% after news that two digital insurers—U.S. auto comparison platform Insurify and Spanish home insurer Tuio—had introduced chatbot assistants using ChatGPT technology. Yet, in the subsequent three weeks, insurance distribution stocks rebounded 7%, surpassing a 1% dip in the broader S&P 500. The market seemed to absorb the AI risk and concluded it was not a substantial danger to revenue growth, adopting a sentiment of “‘nothing to fear’ and ‘far away.'”
BofA holds a contrary view.
“Our perspective is that large-language model digital agents can competently handle a meaningful portion of the work currently done by 20-30 thousand independent agents across the United States,” the BofA report said.
The heart of the firm’s pessimistic argument focuses on a vast number of routine, simple insurance policies. Analysts Joshua Shanker, Joseph Turnillo, Cyril Onyango, and Fatima Keita examined just six major insurers serving small businesses and personal lines: , Hartford, , , Hanover, and Selective. From these six carriers alone, BofA found over $15 billion in commissions paid to independent agents in 2025 that are heavily weighted toward low-complexity risks.
For instance, Progressive paid more than $6 billion to independent agents last year, while Travelers and Hartford paid approximately $3.35 billion and $1.25 billion, respectively, in segments led by personal lines and small commercial business. BofA observes that these policy types, like standard home and auto coverage, involve straightforward transactions where human agents provide minimal added value, rendering direct-to-consumer digital channels a significant potential cost saving for customers.
Amrish Singh, CEO of the AI insurance startup Liberate, told he believes BofA’s estimate is accurate. His own calculations indicate a broad range of $4.8 billion to $33.6 billion in U.S. insurance tasks susceptible to automation.
The snowball effect
While optimists contend that large insurance brokers are not deeply involved in personal lines or small commercial markets, BofA responds that years of ongoing “tuck-in M&A” have produced a “snowball effect.” Hundreds of small acquired agencies have moved a substantial volume of simple, low-premium business onto the books of large brokers, a weakness frequently hidden by inadequate public reporting. Additionally, even large, complex accounts—which are unlikely to be directly bypassed—could see price compression as AI makes insurance markets more transparent for sophisticated corporate clients.
Some investors have compared the AI threat to the heavily promoted but gradual disruption expected from self-driving cars. BofA, however, makes a clear separation. Shifting to autonomous vehicles demands trillions in infrastructure investment and many years, whereas deploying large-language model chatbots is inexpensive, simple, and occurring presently. The report cites Munich Re’s NextInsurance as an example, which already features an AI chatbot on its website allowing customers to directly buy and enact commercial policies without a human agent.
While recognizing that long-term forecasts amid technological change are “difficult,” BofA remarks that Facebook/ and Google/Alphabet did not eliminate print advertising immediately, but over two decades, shifting consumer behavior drastically reduced the print ad market. “We are not arguing that insurance intermediaries will vanish or that will purchase its insurance from a chatbot,” BofA stated, but it advised investors to scrutinize this sector, as insurance distributor stocks do not appear to be pricing in the risks.
BofA highlights that the sector currently trades at 22 times trailing free cash flow and 15 times enterprise value to trailing EBITDA. While bulls may say the stocks appear inexpensive after a 24% fall from peak valuations a year ago, BofA warns these multiples have only reverted to pre-pandemic levels. Moreover, BofA claims insurance distribution companies often employ generous earnings “adjustments”—like excluding integration costs from their continuous acquisitions—that can substantially overstate their genuine earnings capacity.
In the end, BofA is not forecasting the immediate extinction of human insurance agents, nor is it proposing that large corporations like will abruptly procure complex insurance from a chatbot. However, the firm cautions that an agency business currently seen as having 3% to 7% organic revenue growth could see that rate decline to 1% to 5% due to disruptive technology. BofA concludes that with 10% to 20% of present business potentially subject to disintermediation, the industry’s rich valuations allow almost no margin for disappointment.
For this story, journalists employed generative AI as a research aid. An editor confirmed the information’s accuracy prior to publication.