Reports & Research

The Silent Peril Reshaping Homeowners Insurance

Non-weather water damage rarely makes headlines, but it’s quietly eroding profitability across the country.
It is now the fourth costliest peril in homeowners insurance, and claim severity has increased 80% in the last decade—a trend that’s accelerating even as frequency remains relatively flat.

Traditional risk models struggle to capture the early warning signs behind these losses, leading to mispriced policies, undetected exposure, and rising volatility for carriers.

Want the full analysis? Download the complete “Winning the Fight Against Non-Weather Water Losses” guide.

Why Loss Severity Keeps Rising

Aging homes and overlooked system failures

Many of the most expensive losses stem from aging plumbing, deteriorating materials, and slow-burn failures that often go undetected until damage is significant.

Frequency is flat—severity is not

Loss patterns suggest that while the number of events hasn’t surged, the financial impact of each event has—a signal that traditional models are not capturing the right property-level predictors.

The Property Features Most Predictive of Water Losses

The overlooked attributes that traditional models miss

Standard territory- or age-based assessments often ignore the property-specific details that meaningfully influence water loss risk, including:

• supply line material and age
• plumbing configuration
• occupancy patterns
• system maintenance and upgrades
• moisture exposure and prior loss indicators

These factors vary widely between neighboring homes—yet most models treat them as identical.

Where Traditional Underwriting Falls Short

ZIP-code and age-based proxies mask true risk

Legacy approaches rely heavily on broad territory-level assumptions that overlook structural vulnerabilities and system conditions.

Limited visibility creates mispriced policies

Without property-level insight, high-risk homes are often underpriced while lower-risk homes subsidize them—driving loss ratio volatility over time.

Get deeper insights on the drivers of water loss severity in our full guide → “Winning the Fight Against Non-Weather Water Losses”

How AI and Property-Level Data Are Changing the Landscape

AI models trained on real-world claims data can identify early signals of potential water loss by analyzing the interaction between:

• plumbing systems
• property attributes
• historical patterns
• material degradation
• repair history

This enables carriers to segment risk accurately, adjust pricing, and reduce preventable losses—long before small issues turn into major claims.

What Homeowners Actually Understand About Water Risk

Misconceptions around coverage and prevention

ZestyAI’s research shows that many policyholders:

• misunderstand what is and isn’t covered
• underestimate how much damage water can cause
• rarely take preventive actions unless prompted

This disconnect creates an opportunity for carriers to strengthen education, mitigation, and customer engagement.

Steps Carriers Can Take Today

Improve segmentation and rating accuracy

Property-level signals enable more precise risk tiers and more stable long-term portfolios.

Strengthen mitigation and reduce loss severity

Insights help identify which homes are at elevated risk and where targeted mitigation can reduce exposure.

Enhance underwriting workflows with explainable insights

Transparent, explainable AI helps underwriters understand the key drivers behind elevated risk—supporting both decision-making and regulatory review.

Get the Full Guide

Our new research paper, Winning the Fight Against Non-Weather Water Losses, breaks down the trends reshaping this growing peril—and the strategies carriers can use to get ahead of it.

Access the Guide

ZestyAI's analysis revealed that more than 12.6 million U.S. properties are at high risk of hail-related roof damage, representing $189.5 billion in potential replacement costs.

Powered by ZestyAI’s Z-HAIL™ model, the analysis underscores the growing financial threat of severe convective storms (SCS), including hail, tornadoes, and wind events. In 2024 alone, damages from SCS were estimated at $56 billion—surpassing losses from hurricanes.

Yet many insurers still rely on traditional models designed to estimate portfolio-level exposure, not property-level risk. As hail events increase in severity and frequency, these models often miss the structural and environmental conditions that drive real losses.

Kumar Dhuvur, Co-Founder and Chief Product Officer at ZestyAI said:

“Catastrophe models have helped insurers understand where storms may strike and how losses might add up at a portfolio level. But they weren’t built to assess risk at the individual property level, and they often miss the specific conditions that drive hail damage. By analyzing the interaction between structure-specific features and local storm patterns, we can distinguish risk between neighboring properties—enabling smarter underwriting, more precise pricing, and better protection for policyholders.”

Z-HAIL evaluates hail risk using a proprietary blend of climate, aerial, and property-specific data. By applying advanced machine learning to these inputs, Z-HAIL delivers highly granular predictions that reflect both the physical characteristics of a structure and the storm activity in its immediate surroundings.

Key findings from the analysis:

• 12.6 million U.S. structures flagged as high risk for hail-related roof damage
• $189.5 billion in total potential roof replacement exposure

Top five states by dollar exposure:

• Texas ($68B)
• Colorado ($16.7B)
• Illinois ($10.8B)
• North Carolina ($10.4B)
• Missouri ($9.5B)

States with the lowest dollar exposure:

• Maine ($4.7M)
• Idaho ($12.8M)
• New Hampshire ($18.5M)
• Nevada ($49.3M)
• Vermont ($64.7M)

In recent case studies, Z-HAIL has demonstrated the ability to pinpoint which properties are most susceptible to hail damage—even within the same neighborhood and exposed to the same storm. In one example from Allen, Texas, following a storm with 2.5-inch hailstones, Z-HAIL segmented risk across 483 policies, identifying no losses among properties rated “Very Low” by the model. This level of intra-territory precision gives insurers the ability to refine risk selection with confidence—even in the most hail-prone regions of the country.

Severe convective storms are becoming more frequent and costly, putting pressure on insurers to refine underwriting and risk management strategies. 

On April 2, our experts covered:

• Key drivers behind increasing severe storm losses
• What La Niña means for the 2025 season
• How AI-powered risk models improve risk segmentation
• Live Q&A – Get expert answers to your toughest questions!

Missed the live event? Stream now! 

Severe convective storms (SCS)—including tornadoes, hail, and damaging wind events—resulted in $58 billion in insured losses across the U.S in 2024.

Insurers face a dual challenge: navigating the uncertainty of storm patterns while ensuring their portfolios remain resilient enough to absorb the financial strain from clustered, high-loss events.

Research with IBHS confirms that SCS damage accumulates over time, particularly affecting rooftops after multiple exposures to intense storm activity. As housing stock deteriorates, insurers must reassess their portfolios to ensure underwriting, rating, and loss cost controls align with their risk appetite and maintain premiums that accurately reflect evolving exposure.

Get ahead of rising storm risks with expert insights that help you strengthen underwriting, risk assessment, and claims management.

Get our new report.

A staggering $2.15 trillion worth of U.S. residential property is at high risk of wildfire damage, according to a new AI-powered analysis from ZestyAI, the leader in climate and property risk analytics. The study, which assessed 126 million properties nationwide, found that 4.3 million individual homes face heightened wildfire risk—far beyond traditionally recognized high-risk areas.

Using advanced AI models trained on over 2,000 historical wildfires, ZestyAI mapped wildfire exposure at the property level, integrating satellite and aerial imagery, topography, and structure-specific characteristics. While California leads the nation with $1.16 trillion in wildfire-exposed property, other states such as Colorado ($190.5 billion), Utah ($100.3 billion), and North Carolina ($71.2 billion) also face significant risk.

Wildfire Risk is a Nationwide Challenge

While the Western U.S. has historically seen the most severe wildfire activity, ZestyAI’s findings confirm that high-risk properties exist across the country. States like North Carolina (4.6% of homes at high risk), Kentucky (2.9%), Tennessee (2.3%), and even South Dakota (11.0%) are now seeing increased wildfire exposure.

As more homes and businesses are built in fire-prone landscapes, the Wildland-Urban Interface (WUI) continues to expand. This, combined with intensifying climate conditions, is driving higher insurance costs and growing availability concerns. Today, one in eight U.S. homeowners already lacks adequate insurance coverage, and that number is expected to rise.

AI Expands Insurance Access in High-Risk Areas

Attila Toth, Founder and CEO of ZestyAI said:

"Wildfires are threatening more properties than ever before, with billions of dollars in exposure even in areas many people don’t associate with fire risk. Yet, too many homeowners are finding themselves uninsured or underinsured just as these disasters become more frequent and severe. Insurers have traditionally relied on broad, regional models that don’t account for individual property characteristics."

"That means some homeowners are denied coverage even when their true risk is much lower than their neighbors'.’"

AI-driven risk analytics are reshaping the way insurers assess wildfire exposure. By providing granular, property-specific insights, we’re helping insurers make smarter underwriting decisions—keeping coverage available in high-risk areas while ensuring that homeowners who take mitigation steps are recognized.

Last year, our models helped insurers extend coverage to 511,000 properties that had previously struggled to secure insurance due to outdated risk models. In 2025, we expect that number to reach a million, ensuring that even in high-risk areas, responsible homeowners have access to protection when disaster strikes.

Artificial intelligence is reshaping the P&C insurance industry, offering new ways to streamline underwriting, enhance risk management, and navigate evolving regulations.

But as AI adoption accelerates, insurers must ensure they’re using these technologies effectively—balancing innovation with compliance.

Our latest guide explores the most impactful AI applications in insurance, including:

• AI-powered underwriting and predictive analytics
• How regulators are shaping the future of AI in insurance
• Best practices for integrating AI while ensuring fairness and transparency

As AI-driven tools become the new standard, insurers who adapt early will gain a competitive edge.

Download our free guide to leverage these innovations while staying aligned with evolving regulations.

We’re proud to share that ZestyAI has been named to Sønr’s 2025 Beyond Boundaries Scale50, recognizing the top 50 established insurtechs driving measurable impact and transformation across the global insurance industry.

Produced by Sønr, a leading market intelligence firm tracking more than four million companies worldwide, the Beyond Boundaries 2025 report identifies the innovators redefining insurance through AI, data, and collaboration.

This year’s analysis underscores a clear shift in the market: the age of experimentation has given way to execution and scale—where efficiency, resilience, and real-world outcomes define success.

At ZestyAI, we’re proud to be part of that evolution. Our Decision Intelligence Platform brings together property-level data, predictive AI models, and Agentic AI automation to help insurers see, price, and manage risk with precision and confidence.

Trusted by carriers and regulators across the U.S., ZestyAI’s solutions deliver measurable improvements across underwriting, rating, reinsurance, and regulatory workflows—helping insurers make faster and more data-driven decisions.

Matt Connolly, Founder and CEO of Sønr, said:

The insurance industry has long talked about change. And now, we’re seeing it happen. After years of incremental steps, the market is finally embracing the opportunities technology brings - and the impact is tangible.

Read the full report: Beyond Boundaries 2025

DUAL North America Inc.’s (“DUAL”) personal property division has selected ZestyAI’s Z-STORM™ model to enhance storm-risk underwriting and pricing as it continues its rapid US expansion.

The partnership equips DUAL with sharper risk differentiation, more accurate underwriting and pricing, and a stronger foundation for sustainable growth in regions increasingly affected by severe convective storms.

By adopting ZestyAI’s severe convective storm model, DUAL will strengthen its ability to identify and price the combined effects of hail and wind with greater precision. This will enable faster, more informed decisions and profitable expansion while maintaining regulatory compliance.

The collaboration reflects DUAL’s continued investment in advanced analytics and technology to support long-term growth.

The specialty program administrator, offering more than 40 insurance products and surpassing $1.3 billion in gross written premium in 2024, continues to broaden its capabilities across commercial, specialty, and personal lines.

Luke Wolmer, Chief Actuary at DUAL, said:

“As we continue to grow across personal property lines, having accurate risk prediction at the property level is crucial."

Z-STORM gives us a more nuanced understanding of storm vulnerability, helping us recognize differences in risk that traditional models overlook. This enhances our team’s confidence in pricing decisions and will support our continued expansion across the U.S.”

Z-STORM is an AI-powered risk model that evaluates the combined effects of hail and wind to predict the frequency and severity of storm-related damage at the property level. By analyzing the interaction between local climatology and the unique characteristics of every structure—including roof condition, material, and surrounding exposure—the model delivers precise, property-specific insights into storm vulnerability.

In September 2025, ZestyAI introduced mitigation-aware scoring to its severe convective storm suite, allowing insurers to dynamically adjust risk scores to reflect verified improvements such as roof replacements, upgraded materials, or corrected property data. This enhancement gives carriers a scalable way to recognize mitigation within pricing and underwriting workflows, advancing transparency and regulatory alignment.

Attila Toth, Founder and CEO of ZestyAI, said:

“DUAL’s adoption of Z-STORM reflects a forward-thinking approach to storm risk management."

"By applying property-level risk analytics and mitigation-aware scoring, DUAL is positioned to underwrite more precisely, grow responsibly, and strengthen community resilience across the regions that are most exposed to extreme weather”.

ZestyAI’s storm models are regulatory reviewed and ready to use across the Great Plains, Midwest, and U.S. South, regions most impacted by severe convective storms, and are actively used by carriers for rating and underwriting.

ZestyAI today announced the expansion of ZORRO Discover™ to all property and casualty insurance lines.

ZORRO Discover analyzes millions of state filings to surface real-time regulatory and market intelligence, giving carriers actionable insights to make faster, more confident decisions. Carriers using the platform have reduced adverse selection, accelerated regulatory approvals by up to 50%, and expanded analytical capacity more than 20-fold—turning what was once a manual, fragmented process into a source of strategic advantage.

The platform now delivers unified visibility across all P&C lines, including Commercial Auto and Property, Personal Auto and Property, Financial and Specialty Lines, Liability and Professional Lines, Workers’ Compensation, and Administrative filings—covering every major filing type across the United States.

Built on ZestyAI’s Agentic AI platform, ZORRO Discover scales decision intelligence across the insurance industry, transforming over a decade of U.S. insurance filings into a single, transparent system of insight. Carriers can instantly benchmark competitors, analyze rating trends, and anticipate regulator feedback and objections in real time, turning regulatory filings from a compliance requirement into a strategic advantage.

Kumar Dhuvur, Chief Product Officer and Co-Founder at ZestyAI, said:

“Every corner of P&C faces the same challenge: too many filings and too little time. Now, whether it’s workers’ comp in Texas or commercial auto in California, teams can simply ask ZORRO and get instant, verified insights in real time.”

By analyzing past objections and outcomes, teams can anticipate regulators’ questions before they arise and move filings forward with precision. Live monitoring of new submissions keeps organizations current on competitor moves and market shifts, turning what was once a fragmented, manual workflow into a real-time decision system that helps teams act quickly and strategically.

With its conversational interface, users can simply ask ZORRO to surface insights that once took hours or days to uncover. Product, actuarial, and regulatory teams can now collaborate from a single, auditable source of truth, replacing manual searches and static spreadsheets with transparent, explainable intelligence that drives faster, smarter action.

ZORRO Discover is available now for all property and casualty insurance lines.

Start your trial.

‍

At ZestyAI, we know that better data leads to better decisions. That’s why we’ve invested in a major upgrade to our Roof Age product, trusted by leading carriers to improve risk selection, pricing, and operational efficiency in property insurance.

Today, we’re excited to share what’s new, what’s improved, and how these advancements are already helping carriers strengthen underwriting, rating, and inspection workflows.

What’s New in Roof Age

We’ve taken a holistic approach to improving performance, accuracy, and efficiency. Here’s what you’ll find in the latest release:

Refit model with double the training data

We’ve significantly enhanced the Roof Age model, doubling the size of our training dataset to improve performance across diverse housing stock, roof types, and geographies.

This expanded dataset incorporates more confirmed roof replacement events and broader regional variation, allowing the model to generalize more effectively to different parts of the country, including historically underrepresented regions.

The model is now better able to distinguish between full roof replacements and other types of roof-related activity, such as solar panel installation, patched sections, partial replacements, or home additions.

These events may alter the roof’s appearance or condition, but don’t represent a comprehensive replacement. By learning the subtle visual and contextual cues that separate these scenarios, the model delivers more accurate predictions and reduces the risk of misclassification.

Enhanced estimation for challenging cases

In cases where no building permit is available and roof replacement can’t be clearly confirmed via aerial imagery, our improved Roof Age Estimation Model takes over. This model, now trained on double the dataset, is purpose-built for ambiguity.

It leverages not only imagery and property-level features but also regional climatology, using knowledge of local weather patterns and environmental stressors to inform its estimate.

For example, a roof in the Southeast exposed to intense sun and humidity will age differently than one in the Pacific Northwest or Upper Midwest. Incorporating these regional factors helps improve estimation accuracy, even when direct replacement signals are unavailable.

ZestyAI also establishes a minimum roof age, providing additional clarity and confidence. Using our extensive, 20-year aerial imagery catalog, we can identify the earliest visual evidence of the current roof.

If no replacement activity is detected over a known span of time, we can confidently assert that the roof is at least that old.

This minimum age is then used not just as a floor, but as a valuable input to further refine the overall roof age estimate, narrowing the prediction with greater precision than models limited to single-source or snapshot data.

This capability provides underwriters and actuaries with a powerful, high-confidence signal, particularly valuable for pricing segmentation, inspection prioritization, and risk selection strategies.

Intelligent cross-validation logic

The model doesn’t rely on a single data source. Even when a strong signal like a building permit is available, it cross-validates with high-resolution aerial imagery to detect inconsistencies, like permits that were filed but not followed through, or replacements that occurred without permits.

This layered logic helps ensure predictions are grounded in current conditions, not just administrative records. It also improves detection of fraud, data entry errors, or outdated assumptions in property records.

This logic creates a "trust but verify" framework that boosts both precision and confidence in every prediction.

To illustrate, imagine a home built 12 years ago. The model begins by anchoring to the construction year, then scans forward through our aerial imagery catalog and permit records to assess whether a roof replacement has occurred.

By grounding the analysis in the property's timeline, the model avoids misinterpreting the original roof as a newer installation and increases confidence in identifying true replacement events.

Expanded imagery catalog

We’ve enriched our aerial imagery sources to improve roof verification across geographies. The result: more accurate verification of roof replacements and improved model performance in hard-to-cover geographies.

This helps carriers score more properties with higher confidence, especially in rural or previously under-covered regions.

Confidence scores for every prediction

Every Roof Age prediction now comes with a confidence score, helping carriers make more informed decisions. High-confidence predictions can be fast-tracked for automated processing, while lower-confidence scores can trigger secondary review or inspection.

This added transparency empowers carriers to make risk-based decisions not only on the prediction itself, but on how much to rely on it.

Improved Performance Behind the Scenes

We’ve also made significant infrastructure upgrades to enhance product speed and reliability.

• Reduced Latency: Infrastructure improvements have cut average response times to under 2.5 seconds per property, making Roof Age a real-time-ready solution for quoting and policy decisions.
• Stricter Quality Controls: We’ve added new safeguards to filter out imagery that’s blurry, outdated, or contains visual artifacts. Only high-resolution, high-confidence inputs are used to power predictions.
• Scalability: These backend improvements also allow us to handle larger portfolios with more concurrent requests. This is ideal for carriers integrating Roof Age into enterprise systems.

Easier Access for Every Workflow

Roof Age is available wherever you need it:

• In Z-VIEW: Easily visualize Roof Age predictions and supporting evidence with property-level insights directly in our web application.
• Via API: Seamlessly score entire portfolios and integrate directly into your quoting, pricing, or inspection strategies.

Ready to See the Results for Yourself?

The feedback from the market has been tremendous, and we’re just getting started. Want to see the results for yourself? We’re inviting carriers to pilot the new Roof Age model and evaluate its performance on their own book of business.

Get in touch to schedule your Roof Age pilot

ZestyAI announced that Brava Roof Tile, a leader in premium synthetic roofing solutions backed by Golden Gate Capital, has selected ZestyAI to validate the real-world performance of its roofing systems during past storms.

How Brava Roof Tile Uses ZestyAI’s Property and Roof Intelligence

Brava Roof Tile is leveraging three of ZestyAI’s proven solutions to bring greater clarity to roof performance and replacement opportunities. Roof Age synthesizes building permit data with 20+ years of high-resolution aerial imagery, applying advanced machine learning to deliver verified roof age estimates with 97% U.S. coverage.

Within Z-PROPERTY™, Digital Roof applies AI to assess roof complexity, materials, and condition, flagging vulnerabilities before they become costly failures, while Location Insights evaluates the broader parcel to surface risk factors such as vegetation overhang, yard debris, and secondary structure.

Together, these insights provide comprehensive coverage, unmatched accuracy, and fast deployment at scale, turning property-level data into actionable guidance on roof vulnerabilities and replacement opportunities.

Validating Real-World Resilience With Property-Level Data

“Brava is committed to helping homeowners protect their most valuable asset with roofs that combine durability, sustainability, and beauty,” said Matt Pronk, Chief Financial Officer of Brava Roof Tile.

“With ZestyAI, we gain a clear, data-driven view of how roofs perform in the real world and use those insights to guide families toward stronger, longer-lasting protection.”

“Brava Roof Tile is showing how ZestyAI's risk analytics can be applied to validate resilience in the real world,” said Attila Toth, Founder and CEO of ZestyAI.

“Our mission is to protect families, communities, and their financial wellbeing, and our unmatched coverage and accuracy make that possible at scale."

At ZestyAI, we’re often asked:
What Does “Property-Level” Mean? 
What Makes ZestyAI Different from Traditional Risk Models?
Are ZestyAI Models Approved for Use in Underwriting and Rating?

This post answers the most common questions we receive from underwriters, actuaries, regulators, and technology partners—using wildfire, hail, wind, water, and storm perils as examples of how we turn complex data into actionable insights.

What Does “Property-Level” Mean?

Traditional risk models often rely on ZIP codes, territories, or broad regional averages to assess hazard and vulnerability. Stochastic models may support property-specific analysis, but they typically require external data sources, which adds cost, complexity, and inconsistency.

At ZestyAI, we assess each structure based on its physical characteristics and how it interacts with the surrounding environment.

We assess:

• Parcel boundaries and building footprints
• High-resolution aerial and oblique imagery
• Topography, slope, and vegetation
• Structural details like roof shape, materials, and defensible space

By integrating climatology with the built environment, we generate contextual risk scores that capture how each property’s physical characteristics, regional climatology, and historical loss experience interact to shape real-world risk.

That drives smarter decisions in underwriting, pricing, and mitigation, without relying on assumptions or manually sourced data.

What Makes ZestyAI Different from Traditional Risk Models?

Traditional risk tools often rely on:

• Broad hazard zones that can’t distinguish risk within a ZIP code
• Infrequent model updates that fail to reflect current conditions
• Over-simplified proxies—often relying solely on factors like year built—without accounting for deeper structural nuance
• Manual inspections that are slow, inconsistent, and costly

ZestyAI takes a fundamentally different approach.

Our models:

• Use gradient boosted machines that capture complex interactions between property features and environmental conditions
• Are trained on millions of actual insurance claims, not simulations, ensuring outputs reflect real-world loss experience
• Leverage both imagery (e.g., high-resolution aerial and oblique photos) and non-imagery sources (e.g., permits, climatology, topography)
• Continuously incorporate new data to reflect changing exposures
• Provide parcel-level risk scores with full transparency and regulator-ready documentation

With 97%+ U.S. property coverage, ZestyAI delivers national models with localized precision, helping carriers segment risk, price accurately, and respond to today’s evolving climate risks.

Are ZestyAI Models Approved for Use in Underwriting and Rating?

Yes. ZestyAI’s models, including Z-FIRE™, Z-HAIL™, Z-WIND™, Z-WATER™, and Z-STORM™, have been approved for use in underwriting and rating across the U.S.

Our regulatory approach is grounded in a few key principles:

• Transparency: We provide clear, regulator-ready documentation, including model methodology, variable selection rationale, and statistical validation.
• Collaboration: We work directly with carriers and state regulators throughout the filing process, from pre-submission briefings to objections.
• Responsible Innovation: Our models are trained on real-world claims, regularly updated with new data, and built with fairness and explainability in mind.

We support filings with:

• Detailed methodology and input documentation
• Variable importance rankings and validation studies
• Pre-built regulatory summaries to streamline the review
• Ongoing support throughout the regulatory lifecycle

ZestyAI has a track record of success navigating regulatory review. We help carriers adopt cutting-edge risk models with confidence and compliance.

Roof Age vs. Roof Condition: What’s the Difference?

ZestyAI’s models distinguish between roof age and roof condition, treating them as complementary signals that together provide a more accurate picture of roof-related risk, especially for hail and wind.

Roof Age is validated using a combination of building permit data and aerial imagery, analyzed through multiple proprietary methods simultaneously. We assign confidence scores to each roof age and apply minimum roof age rules to avoid false positives, ensuring the data is robust, even in jurisdictions with limited permitting records.

Roof Condition is assessed through computer vision models applied to high-resolution aerial imagery. These models detect visual signs of degradation, such as discoloration, wear, patching, and debris, that may not correlate with official replacement dates.

Why does this matter?

Because many insurers rely solely on reported roof age, which is often missing, outdated, or self-reported.

Our approach captures:

• Properties with older roofs that are still in good shape (and may be lower risk)
• Properties with newer roofs already showing signs of wear (and may be higher risk)
• Up-to-date roof vulnerability that static datasets can’t match

Together, roof age and condition power smarter decisions in underwriting, pricing, and mitigation—grounded in observable reality, not assumptions.

Modeling Approach: Why Gradient-Boosted Machines (GBMs)?

ZestyAI's risk models use gradient boosted machines (GBMs), a machine learning technique that delivers powerful predictive performance while remaining transparent and regulator-ready.

We use GBMs because they:

• Achieve high predictive accuracy by combining many simple decision trees into an ensemble that learns from its own errors over time, ideal for capturing complex insurance risk signals.
• Model non-linear interactions between variables, such as how roof complexity, condition, and regional climatology jointly influence risk, something traditional models or GLMs often miss.
• Enable transparency and explainability, with tools like feature importance rankings, partial dependence plots, and SHAP values that help underwriters, actuaries, and regulators understand what’s driving risk scores.
• Support a wide range of input types, from imagery-derived features to structured data like permits, topography, and property characteristics, all in one unified framework.

The result: a modeling approach that delivers real-world impact, supporting smarter underwriting, better pricing, and confident regulatory adoption.

How Z-FIRE Evaluates Wildfire Risk

Z-FIRE is ZestyAI’s structure-level wildfire risk model, built to capture both traditional wildland fire exposure and the growing threat of urban conflagration. Unlike traditional hazard maps that apply uniform risk zones across ZIP codes or counties, Z-FIRE delivers granular, property-specific risk scores for every structure in the U.S.

The model includes two levels of scoring:

1. Level One: Exposure Risk – Evaluates how likely a structure is to fall within a future wildfire perimeter, based on vegetation, slope, elevation, proximity to the wildland-urban interface (WUI), historical burn patterns, and regional climatology.
2. Level Two: Structure Vulnerability – Assesses how likely that structure is to be damaged if a wildfire occurs nearby. This score factors in structural characteristics like building materials, defensible space, and surrounding fuels, extracted from aerial imagery using computer vision.

Together, these scores provide a more complete view of wildfire risk: not just where fires may happen, but how individual structures are likely to perform.

Z-FIRE also captures non-traditional wildfire scenarios, including embers and wind-driven fires that jump the WUI and ignite dense suburban and urban neighborhoods. This makes the model particularly valuable for identifying concentration risk, urban conflagration, and managing PML across books of business.

Z-FIRE is validated on millions of insurance claims and performs reliably across all geographies—from the forests of California and the grasslands of Texas to emerging risk zones in Colorado, Oregon, and the Eastern U.S.

How Z-HAIL Accounts for Roof Vulnerability

Z-HAIL is a property-specific hail risk model designed to assess not just the likelihood of hail, but how damaging it will be to a specific structure.

Unlike traditional models that rely on historical hail frequency alone, Z-HAIL captures the interaction between local climatology and structural resilience by analyzing:

• Hail climatology: storm frequency, hailstone size, and intensity at a hyperlocal level
• Roof geometry and materials: pitch, complexity, covering type, and other features that influence how hail impacts a roof
• Property-specific vulnerability factors: including building height, exposure, and roof condition (derived from imagery and computer vision)

By modeling how hail behaves in a given location and how a specific roof is likely to perform under those conditions, Z-HAIL delivers precise risk segmentation at the parcel level. 

Carriers using Z-HAIL have seen significant improvements in underwriting performance. In an independent third-party review, Z-HAIL demonstrated a 20× lift in loss ratio segmentation between high- and low-risk properties—enabling more accurate pricing, better risk selection, and actionable mitigation strategies.

How Z-WIND Analyzes Wind-Driven Damage

Z-WIND is a property-specific model that evaluates vulnerability to both straight-line winds and tornadic activity by analyzing how wind climatology interacts with structure-level characteristics.

The model captures:

• Roof geometry: including shape, pitch, and surface area, which influence uplift forces
• Building elevation: to assess exposure to wind at various heights
• Local terrain and land cover: which impact wind speed, turbulence, and exposure to flying debris
• Historical wind climatology: including storm frequency and intensity
• Real-world claims data: to ensure outputs reflect actual loss performance

Z-WIND generates property-level frequency and severity scores, helping insurers move beyond broad wind zones to more precisely identify risk at the structure level. By understanding how specific buildings respond to local wind conditions, Z-WIND enables more accurate pricing, underwriting, and mitigation strategies across both inland and coastal regions.

How Z-WATER Tackles Non-Weather Water Losses

Z-WATER is an AI-powered model that predicts the frequency and severity of non-weather water and freeze claims at the property level, covering every structure in the contiguous U.S. 

While many traditional models depend heavily on basic indicators such as “year built,” Z-WATER combines those inputs with a broader set of property, climate, and infrastructure features to capture the interaction between three core dimensions of risk:

• Construction & Architecture: Property-specific features that influence vulnerability and claim severity, such as number of bathrooms, number of stories, presence of a pool, and overhanging vegetation (a signal for potential tree root intrusion).
• Climatology: Environmental stressors like temperature swings and freeze/thaw cycles that contribute to pipe bursts and system strain.
• Local Infrastructure & Hydrology: How local plumbing systems and electrical grids perform when real-world cold snaps or heat waves exceed what regional codes anticipated, exposing systemic weak points that lead to burst pipes and interior water damage.

These variables are derived from aerial imagery, tax assessment data, and regional climate and infrastructure datasets, all processed through ZestyAI’s proprietary AI framework.

Z-WATER helps insurers:

• Set fair and adequate rates based on true exposure
• Target high-risk homes for mitigation (e.g., water sensors or shutoff valves)
• Streamline operations by automating low-risk decisions and focusing resources where they matter most

What Is Z-STORM?

Z-STORM is a predictive, property-specific model built for carriers that rate hail and wind as a combined peril. It provides structure-level risk scores across the contiguous U.S., enabling more accurate pricing, underwriting, and mitigation decisions.

Unlike traditional territory-based approaches, Z-STORM models how storm climatology interacts with the built environment—capturing the real-world conditions that drive loss at the individual property level.

The model incorporates:

• Storm climatology: frequency and severity of wind and hail events at a hyperlocal scale
• Structural features: roof shape, material, pitch, and condition—key factors in a structure’s vulnerability
• Environmental context: including open terrain and nearby vegetation, which can amplify damage

Z-STORM predicts both:

1. Claim frequency: the likelihood a property will experience a storm-related claim
2. Claim severity: the expected loss as a percentage of Coverage A, providing a more precise view of financial exposure

This dual prediction enables carriers to:

• Accurately rate combined wind and hail risk at the property level
• Target mitigation strategies (e.g., roof improvements that reduce exposure to both hazards)
• Improve risk segmentation and pricing

Z-STORM offers a single, AI-powered solution for capturing the true complexity of convective storm risk—from climate data to construction detail to expected loss outcome.