Hyperexponential, an AI-native pricing platform for property and casualty insurers, is opening a New York office and hiring experienced US insurance leaders as it expands beyond London.
Richard Gunn says the platform prices about $45bn of gross written premium a year, with roughly 30% of that US premium, and Aviva cut model build time by about 75%.
Gunn argues capturing pricing data at the point of decision will become table stakes as insured assets shift from bricks and mortar to intangibles like cyber risk.
Why insurers must stop throwing away their pricing data
Hyperexponential is betting that insurance pricing should be run by actuaries in code, with every pricing decision captured as data. Most insurers still price in Excel and throw the inputs away, which blocks the feedback loop needed to price risks that are increasingly intangible. A platform that keeps models fast to build and data reusable lets carriers react as markets shift, and the company is now taking that pitch to the US.
Takeaways
Insurers discard the data their pricing generates. Many insurers still price a risk in Excel, pull in information, then throw it away. Gunn frames pricing decision intelligence as a tight loop linking data and decisions to outcomes, and says capturing data at the point of pricing will become table stakes.
Actuaries should own and iterate their models. Hx Renew lets actuaries at carriers, reinsurers and MGAs build and deploy their own pricing models in open source languages like Python. Gunn says Excel won because non-technical users controlled it, so change should return to the business.
Aviva cut model build time by 75%. Aviva built 58 models in its first nine months on the platform, and new policy creation fell from about an hour to five or ten minutes. Gunn insists the gains free staff for higher-value work rather than cutting headcount.
Risk is shifting toward intangible assets. Company value has moved from bricks and mortar to intangibles. Insurers now underwrite exposures like Airbnb and cyber, as the CrowdStrike disruption showed. Gunn says the platform often helps clients work with small, sparse, fragmented data sets rather than big data.
Platform data becomes an asset across the stack. Sales start with actuarial teams, but value increasingly reaches underwriting IT and leadership. Hyperexponential aggregates actuarial and underwriting activity into a data asset fed into clients' wider tech stacks, and clients build their own integrations through HxConnect to reduce vendor dependence.
AI used internally before any branding. Hyperexponential avoided slapping an AI badge on its brand but adopted AI quickly in-house. It is releasing prebuilt AI libraries and building a virtual actuarial analyst and an underwriting assistant. A state of pricing report found 91% of insurers invest in AI or plan to.
The US expansion plan
The platform prices about $45bn of gross written premium a year, with about 30% of that premium and about 30% of clients in the US.
Gunn says it suits US excess and surplus lines as well as the London market risks the founders came from.
Hyperexponential just opened a New York office, is building a team of four or five there, and plans a New York user group alongside London.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
It is hiring experienced US insurance leaders to shape US product strategy and will attend ITC Vegas for the first time, October 15 to 18.
Gunn was advised that companies iterate their US operating model five times, so he expects to be wrong at least four times.