AI shopping agents have opened a new front in personal insurance distribution. Small commercial could eventually be next, although agents appear to have a stronger defensive moat than many once expected.
The issue has come into sharper focus since Meta’s Muse pushed AI-driven insurance shopping into the spotlight.
Meta launched Muse in September as a personal AI assistant capable of carrying out tasks on users’ behalf, including shopping for auto and homeowners insurance. Rather than requiring consumers to visit multiple insurer websites, enter their details repeatedly and manually compare quotes, Muse can navigate platforms, retrieve pricing information and help identify competing offers.
Its emergence has raised concerns about disruption to personal lines distribution, particularly if consumers begin using AI agents to automatically shop their policies at renewal rather than staying with their existing insurer.
The industry response has already been divided: Insurify blocked Muse over concerns about incomplete coverage comparisons, while The Zebra announced an integration designed to allow AI assistants to gather policy information and streamline insurance shopping.
Commercial insurance presents a much bigger test. Conning estimates the US small commercial market generated roughly $158 billion in direct premium in 2025. It counts about 36.7 million US businesses with fewer than 50 employees, representing more than 99% of businesses, although the segment stretches from sole proprietors to companies bordering on the middle market.
Automation is already affecting different parts of the market at different speeds.
Research by global investment management firm Conning found carriers are increasingly using third-party data, automated underwriting and straight-through processing for predictable Micro and Core risks, while more complex accounts continue to require human intervention. Insurers are also trying to push that automation boundary further upmarket.
But predictions that technology would substantially marginalize agents have repeatedly failed to materialize.
Even at the lower levels of the small commercial market, which encompasses businesses with few to no employees, “business owners still feel that they have to have their hands held a little bit,” Jay Sarzen, director in Conning's insurance research group, told Insurance Business. “They want to have some level of guidance, some level of expertise.”
Conning’s broader carrier interviews reached a similar conclusion. Independent agents remain the dominant distribution channel, despite the growth of direct digital insurers, embedded insurance platforms and online comparative raters. Executives interviewed for the study said none of those alternatives had overtaken agents as their main route to market.
Still, straightforward professional and home-based businesses are an obvious testing ground for more automated distribution. Conning estimates there are around 31 million non-employer businesses in the US, including more than 23 million it classifies as dedicated business owners. Many have relatively limited coverage needs, while Micro businesses with one to four employees can often be underwritten with little or no human underwriting.
Parts of that journey seem to be beginning. Workers’ compensation specialist carrier EMPLOYERS launched a ChatGPT app this year allowing small businesses to describe their operations conversationally and receive a workers’ compensation premium estimate before proceeding to the carrier’s full quoting process.
Even where automation expands, Sarzen expects brokers to retain an advisory role — although on simpler accounts it may become far lighter.
At the lower end, he described the role as potentially “perfunctory”: reviewing the business, checking that the proposed coverage is broadly appropriate and moving the application forward without extensive risk assessment.
Further upmarket, automation could remove much of the administrative friction that currently consumes agency time. Instead of waiting days for a carrier to produce a bindable quote, technology could allow an agent to return to a client within hours with several options. The broker’s contribution then shifts toward recommending among those choices based on coverage, price, carrier reputation and claims experience.
“That’s where the agent can say, ‘We’ve got three quotes here from X, Y and Z carriers. My recommendation is for you to go with Carrier Y, and here’s why I believe that to be the case,’” Sarzen said.
For complex small commercial risks, that advisory role becomes harder to automate away. Conning’s report found its upper-end “Gap” segment — businesses approaching middle-market complexity — remains more likely to require human underwriting, while carrier executives continue to view strong agency relationships as central to competing in the sector.
And despite years of predictions that small commercial would follow personal auto toward direct distribution, Sarzen remains unconvinced. “We’ve been talking about this for years and saying, ‘It’s right around the corner. It’s going to happen. Who wouldn’t want that kind of purchase experience?’” he said. “But it just hasn’t gotten there.”
