ZestyAI, a provider of an AI-driven risk decision platform for the insurance sector, has published research suggesting that insurers may be overpaying for reinsurance because of shortcomings in the property information they hold, rather than because of the true quality of the risks they underwrite.
The findings come from a live poll of insurance professionals carried out during an InsTech webinar. When asked to name their biggest obstacle in assessing property catastrophe risk, 71% pointed to exposure data that is either incomplete or out of date.
Far fewer chose the other options: 13% cited conflicting views of risk among stakeholders, 11% highlighted the difficulty of verifying data from third parties, and just 3% named a limited ability to tell risks apart within similar portfolios. According to the accompanying whitepaper, 94 people answered this question. In a separate poll on what would make third-party property data more trustworthy, half of participants selected independent validation, while a third opted for proven predictive performance.
In its whitepaper, titled The Uncertainty Premium, ZestyAI contends that missing or unreliable information has real consequences for reinsurance. Where reinsurers are unable to confirm key details about properties in a cedant’s book, they may rely on cautious default settings within catastrophe models. Such assumptions can push up modelled losses, which in turn may influence pricing, attachment points or the amount of capacity on offer.
This creates problems on both sides of the deal. Insurers with well-managed portfolios may not be fully rewarded for their quality, effectively paying for doubt that does not reflect their actual exposure. Meanwhile, reinsurers face the risk of mispricing business because they cannot reliably separate stronger books from weaker ones.
The firm used roof age to show how a single flawed data point can significantly shift a risk assessment. Its research revealed that in 67% of cases, the roof age supplied by policyholders was out by three years or more compared with the verified figure. In 43% of instances, roofs were older than stated by at least three years, which could lead to risk being underestimated. In the remaining 24%, roofs were at least three years younger than reported, meaning lower-risk properties might be priced too high or turned down altogether.
ZestyAI further found that 78% of carriers it reviewed in the Midwest, Great Plains and Southeast link Actual Cash Value (ACV) roof endorsements to roof age, with some thresholds triggered once a roof reaches only eight years old. As a result, a small discrepancy in this one attribute can alter rates, eligibility and the terms of cover.
ZestyAI founder and CEO Attila Toth said, “Reinsurance pricing is only as good as the exposure data behind it, and for most of the industry that data is still self-reported, inspected once and left to age. When reinsurers can’t verify what a carrier knows about its portfolio, they have to price the uncertainty. That means a strong book can end up looking much like a weaker one.
“Everyone in the transaction needs the same complete, verified view of every property. A sharper view of risk benefits both sides, because capital finally flows to the risk that has been verified rather than to the uncertainty around it. ”
Copyright © 2026 FinTech Global
