AI Model to Tackle Home Water Damage Gets Green Light in Five States

SAN FRANCISCO, CA – December 04, 2025 – A new artificial intelligence model designed to predict the risk of indoor water damage—a peril that silently costs homeowners and insurers billions annually—has received regulatory approval in five states, signaling a major shift in how the industry confronts one of its most persistent and costly challenges.

ZestyAI, a company specializing in risk intelligence, announced its Z-WATER™ model was accepted for use in underwriting and rating by regulators in Illinois, Indiana, Iowa, Louisiana, and Wisconsin. The move allows insurers to move beyond outdated, generalized risk assessments and instead price policies based on the specific vulnerabilities of individual properties. This development comes as the insurance market grapples with escalating losses from routine incidents like burst pipes and faulty appliances, which have quietly grown into a catastrophe-scale problem.

The Hidden Catastrophe in the Walls

While hurricanes and wildfires capture headlines, the slow drip of a hidden leak or the sudden rupture of a water heater has become a far more frequent and increasingly severe financial drain. Non-weather water damage now stands as the fourth-costliest peril in homeowners insurance, with the severity of these claims soaring by 80% over the past decade. The average loss from a single incident now exceeds $13,000, and in aggregate, these claims often surpass the financial devastation wrought by major named storms.

For decades, the insurance industry has struggled to accurately predict and price this risk. Traditional underwriting models have relied on broad, imprecise proxies such as a home's age or its ZIP code. This approach creates a system of cross-subsidization, where homeowners in well-maintained properties effectively pay for the higher risks of their neighbors. An older home is not inherently riskier if its plumbing has been updated, just as a newer home is not immune to faulty installation or appliance failure. These traditional methods fail to capture the granular, property-specific factors that truly drive interior water losses, leaving insurers exposed and policyholders with rates that may not reflect their actual risk profile.

This modeling gap has placed immense pressure on carriers' profitability. The inability to differentiate between high-risk and low-risk properties within the same territory makes it difficult to manage portfolio exposure effectively. As losses mount, the only recourse has been to raise premiums across the board, a blunt instrument that penalizes responsible homeowners and does little to address the root causes of the claims.

A Digital Twin for Every Pipe

ZestyAI's Z-WATER™ aims to replace these broad-stroke estimates with surgical precision. By leveraging artificial intelligence and vast datasets, the platform creates a "digital twin" for each property, offering a detailed view of its unique risk landscape. This is a fundamental departure from the old way of doing business.

The technology works by applying computer vision algorithms to high-resolution aerial and satellite imagery. It analyzes features invisible to the naked eye and combines those insights with a wealth of other information, including property-level building records, historical permit data, localized climate patterns, and even the context of surrounding infrastructure. The AI model then synthesizes these billions of data points to understand how different variables interact to influence the likelihood and potential severity of a non-weather water claim.

According to the company, the model can predict the frequency and severity of these claims with up to 18 times greater accuracy than traditional rating tools. This claim is substantiated by training and validating the model against billions of dollars in real, verified loss data provided by insurance carriers. The result is a highly specific risk score for each home, empowering insurers to make far more informed decisions.

"Non-weather water losses place real pressure on carriers' books, but they're also highly preventable when you understand where the risks actually lie," said Bryan Rehor, Director of Regulatory Strategy at ZestyAI, in the company's announcement. "Z-WATER helps insurers pinpoint those vulnerabilities at the property level and price them appropriately, while meeting regulators' expectations for clarity and fairness."

A New Blueprint for Risk and Regulation

The approval of Z-WATER™ in five states is significant not just for its technological innovation, but for what it represents in the broader context of AI adoption in highly regulated industries. Insurance regulators are tasked with protecting consumers and ensuring that rates are fair, not discriminatory, and actuarially sound. Historically, this has made the adoption of complex "black box" AI models challenging.

ZestyAI’s success in securing these approvals—part of more than 80 total regulatory acceptances for its models covering perils like wildfire, hail, and wind—suggests a turning point. It indicates that regulators are becoming more comfortable with AI-driven tools, provided that companies can demonstrate the fairness, transparency, and predictive power of their models. This paves the way for a new era of insurance regulation, one that embraces technology to achieve more accurate and equitable risk assessment.

This trend is not happening in a vacuum. The entire insurtech sector is pushing the boundaries of property risk analysis. Competitors like Hosta.a.i. and Chrp are also using AI-powered platforms to conduct virtual home assessments, identifying risks from corroded pipes to faulty wiring using images submitted by homeowners. This industry-wide shift is accelerating the move away from generalized assumptions toward evidence-based, hyper-local underwriting.

The Ripple Effect on Premiums and Policyholders

For homeowners, this technological revolution will have tangible consequences. The most immediate impact will be on insurance premiums. With property-specific pricing, owners of well-maintained homes or those who invest in preventative measures may see their rates stabilize or even decrease, as they are no longer subsidizing higher-risk properties. Insurers can now offer meaningful discounts for installing smart water sensors or other mitigation technologies that can detect leaks and automatically shut off water flow, preventing catastrophic damage.

Conversely, properties identified by the AI as having a higher risk of water damage may face increased premiums or new coverage requirements. While this may be unwelcome news for some homeowners, it also creates a powerful incentive for proactive maintenance. By identifying specific vulnerabilities, insurers can work with policyholders to mitigate risks before a loss occurs, transforming the traditional, reactive claims process into a more collaborative and preventative partnership.

However, this data-intensive approach also raises important questions about data privacy and algorithmic fairness. As insurers gather more granular data about individual properties, ensuring that this information is used responsibly and without perpetuating existing biases is critical. Consumer advocacy groups and regulators will be watching closely to ensure these powerful new tools are deployed in a way that is equitable and transparent, preventing them from being used to unfairly deny coverage or redline entire communities. The challenge for the industry will be to balance the immense potential of AI to create a fairer, more resilient market with the responsibility to protect consumer interests.