Soteris, a machine learning company backed by Y Combinator, that focuses on the property and casualty insurance market, has come out of stealth with a new artificial intelligence product built to help carriers and managing general agents pull additional profit out of the business they already write.
The company has secured more than $8m in seed funding, led by Spider Capital, with participation from Intact Private Capital, Amplify Partners, DCVC, the Webb Investment Network and Overlook Ventures.
Soteris’s original product, which is designed to sharpen insurers’ loss ratios, has been operating with carriers and MGAs since 2020 and has now scored in excess of 100 million policy submissions, representing more than $180bn in premiums. Growth to date has come largely through founder-led sales rather than marketing spend.
The company is now emerging publicly for the first time to launch a new AI-driven profit optimisation tool, aimed squarely at improving an insurer’s bottom line, alongside results generated through work with its existing client base.
The insurance sector is unusual in that carriers typically don’t know the true cost of the policies they sell (claims losses) until long after underwriting them. Because any single policy either produces a claim or doesn’t, insurers have traditionally had to group policies into broader segments to draw conclusions, treating differences within those segments as statistical noise instead of something that could be forecast. A segment only becomes statistically reliable once it holds enough policies, which means the true performance of any one policy is masked within the average.
That approach carries a real cost. Consultancy McKinsey has estimated that coordinated changes, such as trimming underperforming parts of a portfolio and reclaiming profit from segments that appear healthy on paper, can lift gross underwriting results by 30 to 50%, and its 2025 insurance report identifies deeper segmentation and individualised offerings as the industry’s next direction.
Soteris says its technology was built to tackle exactly that problem. Over more than five years, the company has developed proprietary machine learning models capable of building numerous statistically credible segmentations from a policy history and evaluating them together. Where insurers commonly rely on spreadsheets and pivot tables to produce dozens or hundreds of such analyses, Soteris’s approach can generate millions or billions, depending on the volume of known policy characteristics and the depth of historical data, then combine the results to assess individual policies as a “segment of one.”
Onboarding for an insurer takes less than 90 days, the company said, and once integrated, Soteris returns its assessment via API in under 250 milliseconds, at any stage of a policy’s life, whether it is being quoted, bound or reviewed afterwards.
Customers using Soteris’s original product have typically seen loss ratios improve by five to 15 points within a year of implementation.
Soteris founder and CEO Sunit Shah said, “Every insurer knows they’re writing policies that will lose them money. They just can’t find those policies with the resources currently at their disposal. That’s the blind spot we built Soteris to close. For the first time, an insurer can look at a single policy and know exactly what it’s worth, in time to act on that information.”
Spider Capital partner Minsoo Chi said, “With his rare combination of finance, insurance, and quantitative experience, Dr. Shah is uniquely positioned to crack the toughest mathematical problems in insurance. Soteris already moved analytics from segment averages to policy-level expected loss. Extending that same resolution to predicting profit generation is the natural next step.”
Copyright © 2026 InsurTech Analyst
