Hyperexponential, the AI-native pricing and underwriting platform co-founded by CEO Amrit Santhirasenan, says it handles about 45bn of gross written premium and expects that to nearly double within a couple of years.
Over 30% of Hyperexponential's revenue already comes from America, and one North American client has more than $5bn of premium and 2,500 daily users.
Santhirasenan argues the industry must create better incentives for data sharing, since insurers cannot price emerging risks like cyber from historical data alone.
Why insurance pricing needs infrastructure that remembers its data
Hyperexponential is betting that insurers need pricing infrastructure that captures data and keeps it useful, rather than more pricing models. Insurer data is scattered across old mainframes and modern warehouses alike, and history says little about emerging risks like cyber. A platform that pulls signals from that mess and puts decisions in front of underwriters lets pricing improve over time.
Takeaways
Pricing tools should keep and compound their data. Santhirasenan defines decision intelligence as uniting data with insights and decisions so the data grows more useful over time. Legacy technology captured large volumes and then threw them away. Hyperexponential removes manual database work so clients can focus on algorithms and underwriter decisions.
Insurer data is far messier than banking data. Client data sits everywhere from green-screen mainframes and email to PDFs and Snowflake, even at insurers with $100bn balance sheets. AI is now solving data mapping and flow, which Hyperexponential applies to break down silos. The mess makes open insurance harder than open banking.
History is a weak guide for emerging risks. A pricing actuary must understand the risk and design fair, competitive structures to bear it. For cyber, the past says little. Investor Marcus Ryu, Guidewire's co-founder, calls the work pulling signals out of noise in a world meaningfully different from the past.
Data sharing needs incentives customers can see. Santhirasenan says open banking worked because its benefits were played back to customers. Insurers should be transparent about how data is used without sharing every secret. Customers might share Apple Watch data if they knew it would price life insurance more fairly.
Premium volume is set to nearly double. Clients transact about 45bn of gross written premium, expected to nearly double within a couple of years. More than 30% of revenue comes from America, where one client with over $5bn of premium has 2,500 daily users.
Insurance AI needs its own dedicated attention. AI boosts productivity and could offer a framework for reasoning about risk. Expectations have shifted fast, with users now frustrated below 99% accuracy. Hyperexponential announced virtual agents for underwriting and actuarial work at HxLive, its first conference.
From actuary to founder
Santhirasenan graduated as a software engineer, chose not to pursue engineering, and became an actuary at Catlin, now part of AXA. Former Catlin staff later founded Convex, an early Hyperexponential client.
At Tokio Marine Kiln he built the analytics team from scratch, the first actuary to work at the Kiln Syndicate in its fifty-plus-year history.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
After Tokio Marine merged several entities, he was asked to build a modern pricing platform and found no suitable framework on the market.
That frustration led him to found Hyperexponential, with Hx Renew as its platform. Early on, about 10 people worked from a car park in Southeast London.