What does it take to scale insurance? Mastercard, One Inc, Aquiline and Insurtech Insights on payments, AI agents and claims in New York

What did Mastercard say is holding insurance back?

Mastercard’s Director of Global Insurance opened the morning. He said the idea behind the series was simple: put fintechs, carriers and investors in one room to think about what comes next in insurance and insurance payments, and to speed up the work that follows.

Audrey Rampinelli, Senior Vice President of Risk Management and Insurance Services at Mastercard, gave the keynote. She said insurtech has moved from trying to disrupt and cut out incumbents toward working with carriers, because insurance is complex and heavily regulated. Now the industry sits at another turning point. There is lots of capital and great innovation, she said, but scale is hard to reach.

“What’s missing isn’t ideas. There are tons of really fabulous ideas, but it’s infrastructure, connectivity and execution at scale that’s sustainable.”

Audrey Rampinelli, SVP, Risk Management and Insurance Services, Mastercard

She named three forces coming together. Customers now compare their insurer to the best digital experience they have anywhere. AI and data make real-time underwriting and claims possible. And risk is rising, with fraud made sharper by AI.

Mastercard, she said, embeds insurance in its cards, modernizes payment flows such as instant claim payouts, and connects the wider ecosystem. Her closing point set the tone for the day. Scaling insurance will not come from one company or one funding round. It will come from carriers, insurtechs and investors moving together.

Where does scale break in insurance?

The first panel took on that question head on. Michael Ellis, CFO of One Inc, turned it around. The real question, he said, is “where doesn’t it break?” because it breaks across the whole insurance life cycle, from policy admin and underwriting to claims and payouts.

For Ellis, scale means growth that makes the operating model more efficient, with fewer manual steps and fewer one-off promises the back end cannot support. “It’s no longer a grow at all cost mentality,” he said.

Maria Anthony, Vice President, Ventures at Motive Partners, invests from pre-seed through Series A with a focus on insurance. She looks for two things. Can the founder build repeatable sales that do not depend on them alone? And is the startup solving a real structural problem, or just a short-lived gap in the market? She also warned that many founders do not start their pipeline early enough, given how long insurance sales cycles run.

“Building a great technology product is actually just not enough. You might have built a great product, but you have so much more trust to build in this ecosystem.”

Maria Anthony, Vice President, Ventures, Motive Partners

Nick Seibert of Aquiline said some strong niche businesses will never be the size of a big carrier, and that is fine. He compared early insurtech to early online lending: writing premium is easy, doing it with sound unit economics is not. On broker roll-ups, back-end systems have to be joined up. “Otherwise, all you have is a hundred $5 million businesses, not one $500 million business.”

How should carriers and insurtechs build trust?

Hodgen Mainda, EVP and Global Head of Insurance at Transcard and a former Tennessee insurance commissioner, said the pandemic showed the industry it had to modernize. He listed four gaps: legacy core systems, data, a lack of payment orchestration, and reconciliation. When he talks to carriers, he said, two topics always come up: fraud and AI.

An Allianz innovation leader on the panel compared carrier and insurtech partnerships to a relationship. Carriers slow down, he said, because underwriting and data are the heart of the business, and they will not hand that over before trust is built. He asked startups to be open from the start about funding, runway and timelines, so both sides can move faster on pilots. Mainda agreed that once a carrier signs, the partner has to be ready to deliver at the carrier’s pace.

Looking three to five years out, Mainda predicted more investment in fraud tools and AI, and more partnerships, because “we can’t do this alone.” Ellis said manual reconciliation work will fade while human expertise stays. Seibert offered a contrarian note: some savings carriers hope to gain from AI may be eaten up by plaintiff firms using the same tools to file more lawsuits.

How will AI agents pay for insurance?

Angad Banga, Vice President, Agentic Strategic Business Development at Mastercard, was the first hire on a new team built for agentic commerce. He said business uses, like paying invoices on set terms, may matter as much as consumer shopping.

Mastercard’s starting point is Agent Pay, a tokenization technology that makes it clear to everyone in a payment that an agent was involved. Trust is the core, he said.

“If this thing is not trusted, it will never be adopted.”

Angad Banga, VP, Agentic Strategic Business Development, Mastercard

He also described what he called verifiable intent. If a person tells an agent to make a purchase with set limits, that instruction has to be captured. Otherwise, there is no audit trail when a dispute, fraud case or chargeback comes up. He said agentic payments will need to run on many rails, including cards, bank payments and stablecoins, a point that ties to Mastercard’s agreement, announced in March, to acquire stablecoin firm BVNK.

For insurers, his most practical point was about discovery. If someone asks an AI assistant for renters insurance or rental car cover, the assistant needs good data to answer well, including which benefits their card already carries.

Why are insurance payments still stuck on paper checks?

Mike Kresse, who leads commercial payments and money movement for Mastercard in North America, traced how buying insurance moved from the agent’s desk to websites and now to AI agents. Payments have not kept pace. He said more than half of insurance payments are still made by check.

He was clear that slow payouts are not about insurers gaming the float. The costs of handling checks and fighting check fraud outweigh any float benefit. It is simply inertia. And it matters, because one in three consumers will switch insurers after a bad claims experience. Younger customers do not even know how to write a check, he said, pointing to his own grown daughters.

“All of us know that payment is the moment for these policyholders.”

Mike Kresse, EVP, Commercial and New Payment Flows, North America, Mastercard

How is AI changing insurance operations and claims?

The closing panel, moderated by Lutz Kiesewetter of plenum AG, looked at service and claims. Kristoffer Lundberg, CEO of Insurtech Insights, split AI’s effect in two. The direct effect is doing today’s work faster, like the voice agent Travelers now uses for auto claim calls. The indirect effect is what people can do for customers once that work is automated.

A Mastercard leader for North America B2B payments said insurance has many touch points where real-time payments could help, and that Mastercard is widening how people can be paid, from account-to-account and push-to-card to digital wallets, while making reconciliation easier for back-office teams.

Bobbie Shrivastav, Co-Founder and CEO of Solvrays, shared why she builds in insurance. When she became a mother, her maternity benefit was tied to a disability product, and her only “digital” option was a fax machine. The real bottleneck, she said, is fragmentation: so much work is still stuck in emails, spreadsheets and legacy systems. Her advice: start with real business processes and bring staff along.

Robin Roberson, co-founder of WeGoLook and Agentech, said earlier waves of claims tech were rules-based and brittle. Agentic AI can now read messy documents, grainy photos and handwritten notes, and run whole workflows. But she drew a clear line on judgment.

“On these non-judgment tasks, that’s where I would get really bullish, aggressive and unapologetic about using AI in your organization.”

Robin Roberson, serial insurtech founder

For a total loss or a claim denial, she said, people need to step in with empathy and context. Izik Lavy, CEO of GeoX, said its aerial and satellite data shows the worst roofs carry far higher claim frequency. Dr. Robin Kiera of Digitalscouting said insurers exploring digital assets should work through trusted payment and banking partners, because trust is the industry’s oldest strength.

Lundberg closed with two shifts. Teams often think fast and execute slowly; the aim should be to think slowly, then execute fast. And AI itself is now a risk to insure. AI jumped from tenth to second place, behind cyber, in the 2026 Allianz Risk Barometer. When agents talk to each other and make payments, he asked, who takes the liability, and how do insurers underwrite it?

Topics