Last year, 60% of the world's losses went uninsured. That’s a $9 trillion problem, considering the value of the risk involved. But it’s not because insurers are bad at pricing risk. They're just bad at affording to service it, especially for smaller accounts. There’s another problem, too: the industry still runs on faxes, legacy portals, and phone calls, so insurance isn't where you'd expect AI to break through. As a result, most software companies keep their distance. Pace didn't. The company just raised a $46 million Series B.
Founder and CEO Jamie Cuffe has a direct line to the industry's biggest pain points. He grew up around his father's work in reinsurance and brokerage, then went on to found Retool, a platform that enabled companies to build internal business software without writing it from scratch. Now, that mix of insurance fluency and product instinct is showing up in Pace’s traction: since launching last year, Pace agents have completed more than 250,000 critical workflows, growing 3x every quarter, for insurers including Prudential, Newfront, and Palomar. At Prudential alone, Pace is automating thousands of hours of policy servicing and issuance work.
Behind those numbers is a simple design choice. Pace agents do the work end-to-end: submission intake, policy servicing, claims handling, data entry. Rather than wait for insurers to modernize their systems, the agents navigate the systems as they already exist, clicking through internal apps and reasoning across documents.
Lessons from the front lines
Pace is already sharpening how we think about AI-native services.
Domain credibility opens doors that cold outreach never will. Cuffe's early conversations with Prudential landed because his background gave him instant fluency in the room. Those conversations turned into deployments, and these early deployments became the springboard for the marquee carriers that followed.
There's also a version of what economists call the Jevons Paradox playing out in how Pace handles quality assurance. The original observation, from 19th century economist William Stanley Jevons, was that more efficient steam engines didn't reduce coal consumption, they increased it, because cheaper power created new uses for coal that hadn't been worth pursuing before. Insurance operations are showing the same pattern.
Make a service dramatically faster and cheaper, and buyers use more of it. Traditional BPOs QA a sliver of completed work because human review is expensive. Pace can effectively review all of it, because the constraint that made sampling necessary is gone. Fewer than ten people at Pace handle work Cuffe estimates would take close to a thousand under a traditional outsourcing model. Cost reduction is the headline. But the real story is how much the total addressable market opens up once quality assurance stops competing with scale.
Underpinning Pace's success is an advantage shared by many AI-native services: they become more capable as the underlying LLMs improve. Cuffe has pointed to a single model release earlier this year that took Pace's success rate navigating legacy insurance interfaces from around 30% to over 95%.
We're proud to back Jamie and the Pace team as they take on a problem that's structural rather than incidental. When servicing costs stop being the reason coverage doesn't exist, more of the world's risk gets insured.
