Insurance carriers among the Fortune 100 are still losing ground to a small commercial MGA on the one piece of technology increasingly indefensible to distribution partners: a working API.
“Hearing their perspective, saying they’ve been trying to solve that for two years and they’re still so far behind where you’re at today – that just kind of shows the mindset and how long it takes to change in this industry,” said Mark Seich, chief revenue officer at Coterie Insurance, a tech-enabled, data-powered managing general agent (MGA) specializing in small commercial property and casualty coverage. The observation came directly from new hires on his distribution team who came from those large national carriers. PwC research has found insurers spend an average of 70% of their annual IT budget simply maintaining existing systems, leaving little room to build the API capability distribution partners now expect.
Opposed to a want of resources, the gap exists due to engineering speed and flexibility; the ability to adapt an integration to what a distribution partner’s own systems need, rather than offering a single off-the-shelf connection, said Seich. "Is it an off-the-shelf solution? How fast can they be able to accept what it is that you need and deliver for you? Or are they telling you it’s this box and this box only?" he said. "I think that’s the difference when you’re really evaluating that in today’s environment."
Coterie is an MGA, but Seich said it operates more like a full-stack carrier – which flips the usual evaluation dynamic. Distribution partners come to Coterie already committed to consuming an API and already clear on what they need from it, putting Coterie on the hook to prove it can deliver. "They’re usually challenging the carrier," Seich said, "to prove it can support their model, whatever that model looks like." The stakes of that flip are rising with the MGA channel itself: MGA-produced premium now represents roughly 12.5% of total U.S. P&C premium, according to Gallagher Re, up about 10% in 2025 alone.
The criteria distribution partners weigh are consistent regardless of carrier size; A vendor should be judged on results, not promises, said Seich. Key metrics include speed-to-price, submission-to-quote, and the ease of automation between such processes. The stakes of getting that wrong are measurable: only 55% of small commercial customers said in 2025 they "definitely will" renew with their current insurer, down six points from a year earlier, according to J.D. Power, with service outweighing price as the bigger retention driver.
Not every integration carries the same weight, Seich said. At Coterie, clients write commercial insurance large and small. "If you’re trying to API and integrate large commercial, that becomes much more complex because of the unique needs that are needing to happen there." Excess and surplus lines and auto sit further along that curve, where "API calls and integrations become a little more intertwined." A professional liability or general liability policy, by contrast, "is similar to a homeowner’s policy integration." The pattern matches the broader market: McKinsey research indicates up to 95% of personal-lines policies at leading carriers already move through straight-through processing, versus an expected 2030 timeline for similar automation in small commercial.
Distribution partners also rarely want the same shape of connection. Carriers that struggle to offer solutions often offer a single integration shape and ask every partner to conform to it; the ones that succeed flex to meet each partner where they are. That matters because commercial premium still doesn’t flow through one channel: independent agents wrote 87.7% of U.S. commercial lines premium in 2025, per the Independent Insurance Agents & Brokers of America, meaning an API built only for direct or embedded distribution misses most of the business.
That same principle, building the capability before layering something new on top, extends to AI. "We were already doing kind of the data integrations that way," Seich said. "All we did was start to turn and let AI start to do some of it for us." He contrasted that with a common mistake: "Some competitors are trying to bring that data in and do it with AI at the same time. It inflates what the value-add of AI was, because they never did step one. They went right to step five."
For C-suite technology buyers, that distinction matters beyond any single vendor. A fast quote engine or an AI layer sitting on an unintegrated data pipeline won't perform the way a sales demo suggests. The more durable signal is whether speed and conversion metrics can be proven against a real data foundation, not whether the pitch looks impressive.
