YC-backed insurtech Soteris has launched an AI profit acceleration platform designed to help P&C insurers identify profitable and loss-making policies inside their existing books.
The Richmond, Virginia-based company says the product targets a blind spot in traditional insurance analytics: carriers often understand profitability at segment level but lack equivalent visibility into individual policies.
Soteris has operated with carriers and MGAs since 2020. Its first product focuses on expected loss ratios and has scored more than 100 mn policy submissions representing over $180bn in premiums.
The company developed much of that business through founder-led sales without advertising expenditure. Soteris is now emerging publicly with a second product focused directly on policy-level profit contribution.
Soteris has raised more than $8 mn in seed funding. Spider Capital led the financing, with participation from Intact Private Capital and Amplify Partners. DCVC also invested, alongside the Webb Investment Network and Overlook Ventures.
Property and casualty insurers face an unusual accounting problem because claim costs become clear only after policies have been sold. Individual policies also produce uneven outcomes, since one risk might generate a claim while another doesn’t.
Insurers have traditionally dealt with this uncertainty through portfolio segmentation. Actuaries group policies sharing similar characteristics, then evaluate performance across the combined population. Those groups need sufficient volume before actuaries consider the resulting data statistically credible. Individual policy economics therefore tend to disappear inside segment averages.
Soteris argues this approach leaves meaningful differences between policies untreated. The company built its machine learning models around identifying smaller patterns inside large insurance datasets rather than relying solely on conventional segmentation.
McKinsey has previously estimated gross underwriting improvements of 30% to 50% through coordinated actions including portfolio pruning and recovering profitability within apparently healthy segments.
Its 2025 insurance report also pointed toward more granular segmentation and increasingly individualized insurance propositions.
Soteris has spent more than five years developing proprietary machine learning models intended to analyse policy performance at greater resolution.
Traditional insurance analysis often relies on spreadsheets and pivot tables. Such methods might produce dozens or hundreds of credible segment combinations depending on portfolio size and available variables.
Soteris says its methodology produces millions or billions of potential segmentations when enough policy characteristics and historical data exist.
The platform then examines intersections between those analyses to estimate expected performance for individual policies. Soteris describes the result as treating each policy like a credible segment of one. Implementation takes less than 90 days, according to the company.
Once deployed, Soteris delivers results through an API in under 250 milliseconds. Insurers receive the information during quoting or binding, as well as later during the policy lifecycle. The company’s original product estimates expected loss ratios at policy level.
Soteris says customers using the product have recorded loss ratio improvements of 5 to 15 points within less than one year. Those results led the company to examine another problem appearing in customer portfolios.
Insurance economics often pass through several organizations. One company might sell and service a policy, another might hold the required state licence, while a separate balance sheet provider supplies underwriting capital.
Each participant receives a different share of the economics. That makes policy-level profitability harder to estimate using loss expectations alone.
The platform estimates the contribution an individual policy makes to the insurer writing it, rather than stopping at expected claims performance.
Founder and CEO Sunit Shah said insurers already know some policies inside their portfolios will lose money. The harder problem is identifying them early enough for underwriters or portfolio managers to respond.
Soteris gives insurers a policy-level estimate of economic value before those results emerge through historical claims experience. The company has tested its profit acceleration product through several insurer proofs of concept.
Soteris reports book-level EBITDA increases ranging between 70% and 125% in those exercises. The company says those results indicate insurers have substantial unrealized profit variation inside portfolios already on their books.
Shah said insurers gaining better visibility into existing policy economics could redirect additional profit into operations, pricing or customer experience.
Spider Capital partner Minsoo Chi said Shah’s background across finance, insurance and quantitative analysis informed the company’s approach. Chi said Soteris first moved its analytics beyond segment averages toward policy-level expected losses. Extending the same resolution into estimated profit contribution represents the next stage of the company’s product development.
Soteris positions its platform around several stages of the policy lifecycle. Insurers receive policy-level metrics at quote, endorsement and renewal.
Those metrics include estimated profitability and expected loss ratio. The system identifies negative-profit policies alongside risks with high expected loss ratios that remain hidden inside broader segment averages.
The technology also identifies policies with stronger expected economics, giving insurers another way to examine portfolio composition without changing rates, forms or regulatory filings.
Soteris says its new product works within an insurer’s existing distribution network and doesn’t require additional headcount.
The company joined Y Combinator’s 2019 cohort. Its investors include Spider Capital and Intact Private Capital, along with Amplify Partners and DCVC.
The Webb Investment Network and Overlook Ventures also back the company. Clocktower Ventures and Foundation Capital are among its other investors.
