When a major storm headed toward Florida, an insurance claims servicer called Ryze Claim Solutions had the obvious problem: thousands of new claims, one week, not enough people.

The old answer was to find a small army of temporary claims processors and hope the paperwork did not drown them. The newer answer, according to Pace founder and CEO Jamie Cuffe, was to send in AI agents.

"We were able to process thousands of claims immediately," says Jamie Cuffe, the 29-year-old founder and CEO of Pace. "Instead of having a backlog, they had zero backlog. They were able to get through these claims immediately.” The AI agents pulled it off with over 99% accuracy, he added.

Cuffe grew up in the insurance industry. His father ran operations for a reinsurer in Bermuda and later an insurance broker in London, giving him intimate familiarity with the market's operational pain points.

Today, the New York City-based startup said it raised $46 million in Series B funding co-led by Thrive Capital and Sequoia Capital valuing the 28 person company at $375 million.

Insurance is a huge market with a very old plumbing problem. The industry accounts for over 11.8% of U.S. GDP, and about 25 cents of every premium dollar goes to administrative processing. Historically, carriers lowered this overhead by outsourcing to massive offshore human call centers in places like India and the Philippines, run by firms like Cognizant and Wipro.

Pace’s bet is that the next labor pool is not offshore. It is virtual, elastic and indifferent to weather events.

Back-office insurance industry operations are hitting a technological "tipping point" driven entirely by model capabilities, said Lauren Reeder, an investor and partner at Sequoia Capital.

"In different verticals, once the models are good enough to do maybe 50, 60, 70% of a task, the whole vertical very quickly switches and starts to adopt AI very quickly," she said. "We saw this first in legal, and then we saw this in customer support. And then we see insurance as this next wave."

The technical catalyst are recent breakthroughs in front-end "computer use”, or AI models that can click around to control a computer’s interface like a human can. Because insurance relies on archaic legacy portals, faxes, phone calls and PDFs, building custom APIs is logistically impossible. Pace's agents mimic human behavior to navigate these non-API systems end-to-end.

"Our evaluation success rates for navigating legacy interfaces shot up from 30% to over 95% pretty much in the course of one model release at the beginning of this year," Cuffe says.

Cuffe said Pace works with insurance giants like Prudential, Willis Towers Watson, Convex, and Palomar. In some cases, its AI agents can autonomously resolve 90% of policy servicing cases over email and voice, he said.

That is the business case investors are underwriting: a replacement for fixed back-office capacity with software that can swell during a crisis and shrink when demand falls.

"Pace does the exact same thing for knowledge work in insurance that AWS did for cloud computing,” said Thrive Capital partner Philip Clark. “If a severe weather event hits... a carrier can instantly spin up a thousand virtual, highly trained intelligent workers for a day to handle the influx, and then scale them back down when the sun comes out."