What Keeps Insurtech AI Stuck in Pilot

Zinnia CEO George Esposito sat down with Digital Insurance for its Meet the Insurtech Leader series to talk about where life and annuities technology goes next. His answer is less about adding new tools and more about connecting the ones the industry already has, and about building the foundation that lets AI leave the pilot stage.

In the Q&A, Esposito covers his path from corporate law to carrier operations to the CEO seat, how AI will change the economics of a paper-heavy industry, and the mistake he sees insurtechs make most often.

Esposito has worked on nearly every side of the business: as a corporate lawyer at Milbank, as deputy general counsel at Nassau Financial Group, and as president and COO of Everly Life, a direct-to-consumer carrier he helped build. He joined Zinnia in 2021, led the acquisitions of Policygenius and Ebix's life and annuity business, and became CEO in November 2025.

It was at Nassau, mapping the core systems of a 100-year-old insurer, that he saw how far life and annuities had fallen behind.

Asked how AI will change insurance over the next five years, Esposito pointed to the paper that still moves through life and annuities operations. Zinnia sees millions of pieces of it a year, and is building agentic tools that read it, digitize it and link it to the right policy, work that used to be manual data entry.

Removing that friction, he said, means faster answers for advisors, fewer errors for carriers and a better experience for the end customer, and it lets specialists spend their time on the work that needs them. He also set a concrete bar for the buying experience: “By 2030, the buying experience for an annuity should be as simple as opening a CD. That's not a moonshot. We're building it now.”

For the next six months, Esposito sees the biggest opportunity in joining up the pieces of the insurance journey that already exist in digital form, rather than adding more point solutions on top of fragmented workflows.

Distribution is his clearest example. Advisors still quote in one system, run illustrations in another and submit the application somewhere else. Zinnia's connected distribution experience in Zinnia Market Connect brings quoting, illustrations, order entry and case management into one system, which distributors can plug into however they have built their own front end.

The biggest mistake Esposito sees among insurtechs is “shipping an AI demo and calling it a strategy.” Because everyone has access to the same models, he argues, a flashy pilot sets no one apart. What does is the unglamorous work underneath: proprietary data, a real system of record and governance you can audit.

In his view, the winners over the next few years won't be the companies with the best model, but those with the best foundation under it, so that their AI is something clients rely on every day.

Looking ahead, Esposito described the goal as self-serve across the business. For most carriers, getting a new product to market still takes a project, an engineering team and a build cycle, with the distance from launch to a customer's hands measured in quarters or years. “It should be weeks,” he said. Zinnia Launch and Zahara are built to turn a new product into configuration a carrier does itself.

On the distribution side, Zinnia Market Connect is rolling out both as a standard set of enterprise APIs for distributors who want full control and as a white-label, end-to-end platform for those who want a faster path. Underneath it all, Zinnia will keep putting agentic AI to work across its operations so teams spend less time rekeying data and more time serving clients.

Esposito also shared leadership lessons from his career, including a piece of advice to his younger self: “Make the call sooner.”