Resources

Wildfire dominates the headlines, but the everyday fire — the cooking accident, the electrical short, the dryer that overheats — drives more than 20% of all U.S. property claims and 22% of every dollar paid out in homeowners insurance. A new ZestyAI on-demand session examines why everyday fire (also called non-weather fire) has become the least understood and most expensive peril in homeowners portfolios — and introduces Z-SPARK™, ZestyAI's new property-level model for predicting fire risk before ignition.

About this session. Everyday Fire Risk: The Hidden Severity Problem in Homeowners is an on-demand webinar covering why non-weather fire severity has risen 43% in four years, what ignition signals traditional underwriting misses, and how Z-SPARK turns property-level data into actionable claim frequency and severity scores. Presented by Abdul Mohammed (Product Marketing) and Alex Kallos (Director of Product) at ZestyAI.

Prefer to watch instead? Access the full on-demand session → — includes a live Z-SPARK demo and Q&A.

Why is everyday fire the largest hidden risk in homeowners insurance?

Three facts that don't usually appear together:

• Non-weather fire generated $25 billion in insured losses in 2024, or 22% of every dollar paid out in homeowners insurance that year.
• 26 cents of every premium dollar in a representative sample of carriers goes to non-weather fire — making it the single largest base-rate component, bigger than hail and wind, and roughly four times the size of hurricane.
• Over the past five years, wildfire destroyed roughly 35,000 structures. Non-weather fire produced 1.7 million incidents — a 50x volume difference.

Everyday fire isn't hiding because it's small. It's hiding because it's been the least systematically measured peril in the industry — assessed primarily through community-level fire protection (the nearest fire station, the response time) rather than at the individual property.

Why has fire claim severity jumped 43% in four years?

ZestyAI's analysis shows average non-weather fire claim severity climbed from $120K in 2020 to $173K in 2024 — a 43% increase, while claim frequency stayed essentially flat at 0.15–0.16%. This is a severity problem, not a frequency problem. Three drivers explain it:

• The escape window collapsed. UL Fire Safety Research Institute data shows the time between smoke alarm activation and untenable conditions has shrunk from 17 minutes (40 years ago) to 3 minutes today — driven by synthetic furnishings, open floor plans, and lightweight construction. Meanwhile, fire department response averages 7 minutes against a 4-minute national standard. Many fires reach the whole house before help arrives.
• Rebuild costs spiked. Building material prices are up 40% over five years; construction wages up 20%. Add stricter code compliance and longer rebuild timelines and a moderate claim five years ago is a major loss today.
• Smaller fires hide. A single fire claim can raise a homeowner's premium by 29%; a second by 60%. Some homeowners absorb the cost rather than file. Others file and switch carriers. Either way, the carrier writing the next policy inherits a property whose early-warning history is invisible — and the next claim is usually larger.

Once ignition happens, the loss is mostly set. The only meaningful place to intervene is before the fire starts.

What ignition signals do traditional fire risk models miss?

Four that show up repeatedly in ZestyAI's analysis:

• The vacancy next door. A vacant building within 10 meters of a property increases fire risk by 25%.
• The five-foot ignition zone. The first five feet around a structure is the most critical area for ignition. Combustibles in that zone can be the difference between containment and total loss.
• Deferred maintenance. A ZestyAI homeowner survey (500+ respondents) found that roughly half don't regularly clear combustible debris around their property.
• Battery-only smoke detectors. 65% of fatal fire incidents involve battery-only alarms rather than hardwired systems.

All four are predictive. None is consistently captured in current underwriting.

How does Z-SPARK change everyday fire risk assessment?

Z-SPARK is ZestyAI's property-level non-weather fire model, built on millions of fire incidents and hundreds of thousands of verified claims from national carriers. For every property it returns two 1-to-10 scores — Claim Frequency and Claim Severity — along with the top risk drivers behind each score, surfaced automatically so underwriters, customers, and regulators all see the why behind the number.

The segmentation power matters. Across Z-SPARK's hundred discrete risk segments, the highest-risk tier shows roughly 30x the measurable risk of the lowest tier. Two homes next door to each other in the same neighborhood can produce wildly different scores based on debris accumulation, vegetation density, and maintenance state — a level of discrimination that territory-based assessment can't see.

Z-SPARK sits alongside ZestyAI's wildfire model (Z-FIRE), which has received more than 200 regulatory approvals across 41 states. The non-weather fire model builds on the same foundation but turns the lens to the peril carriers were never able to assess at the property level.

What can carriers do with property-level fire intelligence?

Four moves become available once everyday fire risk is visible at the property: identify the highest-loss properties already sitting on the books, prevent new ones from entering as misclassified low-risk policies, segment to fast-track clean risks while setting guardrails on the rest, and allocate inspection resources to the properties that actually need them. Expansion into historically high-risk geographies also becomes tractable, because the risk is no longer a black box at the territory level.

The everyday fire risk hiding in a homeowners book doesn't have to stay hidden. Once it becomes visible at the property level, the entire economics of the peril change.

Watch the full session on demand

Everyday Fire Risk: The Hidden Severity Problem in Homeowners →

Featuring Abdul Mohammed (Product Marketing) and Alex Kallos (Director of Product) at ZestyAI, the session walks through the severity drivers behind the 43% jump in average claim size, the four ignition signals traditional models miss, a live Z-SPARK demo, and Q&A on residential vs. commercial coverage, model explainability, claims-data bias, and how to combine frequency and severity scores in pricing and inspection triage.

Watch the on-demand session — or request a Z-SPARK walkthrough to see how property-level non-weather fire scores apply to your own book.

Across the P&C industry, delayed rate approvals are costing an estimated $72.8 million per day in foregone premium — and the largest single category of objection causing those delays is procedural, not substantive. A P&C Specialist article published May 13, 2026 by Jennifer Ortakales Dawkins covers a ZestyAI analysis of more than 2 million P&C rate and form filings on SERFF, with commentary from ZestyAI's senior director of regulatory and government affairs Bryan Rehor on what's actually driving objection cycles and how filing teams can systematically reduce them.

Read the full article in P&C Specialist →

How much does P&C rate filing delay actually cost?

About $72.8 million a day across all lines, per the ZestyAI analysis. That figure is the aggregate cost of delayed approvals translating into lost premium — premium the rate change would have produced if it had taken effect when intended rather than weeks or months later. The cost compounds two ways: directly, in foregone premium for the period of delay, and indirectly, in continued exposure to the loss patterns the rate change was meant to address.

It also lands unevenly. California took a median of 267 days to approve a rate filing in 2025; Maryland took 206. At the other end of the spectrum, use-and-file states like Wisconsin clear filings in a median of three days, and Wyoming doesn't require rate filing at all. Most of the daily-cost burden concentrates in the slow-approval states.

What causes most rate filing objections?

According to the ZestyAI analysis, 74% of carriers receive objections on at least half of their filings. The most common reasons are procedural rather than substantive: submission gaps, missing or inconsistent supporting exhibits, unclear rationale for the rate change, and incomplete responses to regulator inquiries.

"Submission gaps are a very common and avoidable cause of delay," Rehor told P&C Specialist. The deeper challenge is that each state has its own filing requirements, and the volume of state-specific procedural rules makes it easy to miss something even when the substantive content of the filing is sound.

Which states have the highest rate filing objection rates?

For homeowners filings, the top objection rates are concentrated in the Northeast and the largest markets: New Jersey (87.7%), Massachusetts (87%), New York (84.5%), California (83.1%), and Texas (80.9%). For auto, California (87%) and Massachusetts (85.2%) lead, followed by New Jersey, Texas, Kansas, and Michigan.

Texas has the highest absolute number of objections in both lines, but its overall objection rate is moderated by very high filing volume — more than double the second-place state in either line. The most common reasons for objections in Texas were underwriting errors in home filings and missing or incorrect values in auto filings.

Why does the second objection round matter so much?

Because delay compounds, not stacks linearly. ZestyAI's data shows that after the first objection, each additional challenge adds approximately two months to the filing process as review clocks reset and the scope of scrutiny expands. That's why preventing the first objection is worth more than resolving it efficiently. Filing teams that systematically reduce procedural gaps before submission collapse the timeline far more than teams that respond well to objections after they arrive.

What separates fast-moving filing teams from slow ones?

Less about regulatory environment than execution. The P&C Specialist article includes practical guidance from state insurance departments — a Pennsylvania regulator's reminder that carriers should actually use the department's checklist before submitting, and a Washington regulator's note that subjective language in rate manuals ("above average," "better") will get flagged because regulators require any two people reading the manual to arrive at the same premium for the same risk.

The unifying point: filing teams that internalize the procedural patterns regulators care about — checklists, supporting exhibits, specific language, complete responses — systematically reduce both objection volume and the number of objection rounds. That's the operational gap behind the $72.8M-a-day cost. It's mostly addressable.

Read the full article

The Industry's $73M-a-Day Problem: Rate Filing Delays →

P&C Specialist, May 13, 2026, by Jennifer Ortakales Dawkins. Includes state-by-state objection data from ZestyAI's analysis, additional commentary from regulators in Pennsylvania and Washington and actuarial consultants at Perr&Knight, and detailed examples of what drives objection cycles in the slowest-approval states.

AI-driven property models in U.S. home insurance are shifting from experimental tools to a baseline capability for underwriting — and the economics of consecutive record catastrophe years are accelerating the move. P&C Specialist reporter Vrushank Nayak detailed the shift in a May 2026 analysis, Ignore AI Property Models at Your Own Risk?, featuring perspectives from ZestyAI co-founder and chief product officer Kumar Dhuvur and senior director of regulatory & government affairs Bryan Rehor.

Read the full article in P&C Specialist →

What's actually driving the shift?

Per Dhuvur, the pressure is economic, not technological. The U.S. property insurance market absorbed roughly $89B in insured natural catastrophe losses in 2025 and $108B in 2024 per Swiss Re Institute — what Dhuvur described as "the most expensive operating environment in the history of U.S. property insurance," with no visible path back to historic loss levels. In that environment, the territory-level averages and patchy property-level inputs underwriters have historically relied on aren't fine enough to differentiate risk anymore.

What changed on the supply side is just as important. Aerial and satellite imagery, combined with computer vision, now generate consistent structure-level signals at portfolio scale. Roof condition, structural characteristics, visible damage, and prior exposure can be evaluated across an entire book in the same way they'd be assessed on a single inspected property. That capability didn't exist seven or eight years ago.

How widespread is AI property model adoption today?

Dhuvur estimates roughly half of U.S. carriers now use AI property models in some form, with underwriting the most common use case. Pricing adoption is slower, in part because of regulatory complexity. The P&C Specialist analysis of state rate filings identifies major national carriers — Nationwide, Liberty Mutual, Allstate, and State Farm — among those citing third-party AI property models in homeowners filings, alongside specialty and regional carriers using a wider mix of providers.

Why don't rate filings always reflect actual AI model usage?

This is where Bryan Rehor's regulatory perspective in the article matters most. Filing citation counts can understate real adoption because filing rules differ state to state. Some states require carriers to resubmit their full rating manuals with each filing; others require only the portions that have changed. In change-only states, carriers may continue using an AI model year over year without re-citing it in every filing. Colorado, for example, doesn't require the AI model to be mentioned in every rate filing, but does require carriers to document and govern any model in use, because regulators can audit at any time.

The implication: market-wide AI property model adoption is harder to read from filings alone than the raw citation counts suggest.

What's the cost of moving slowly?

Dhuvur framed the strategic risk as a compounding adverse-selection problem. When some carriers price properties at their true individual risk and others continue pricing on territory averages, the precise pricers systematically capture the better risks and the territory pricers absorb the worse ones. Over time, that gap compounds into loss ratios, retention, and growth — and the carriers most exposed are the ones whose competitors have already moved.

As AI property models shift from differentiator to default, the question isn't whether to adopt. It's how to govern adoption credibly enough to satisfy regulators while still capturing the underwriting gains.

Read the full article

Ignore AI Property Models at Your Own Risk? →

P&C Specialist, May 4, 2026, by Vrushank Nayak. Includes additional commentary from Patrick Schmid at the Insurance Information Institute and other experts, plus detailed analysis of which carriers and states are citing AI property models in current rate filings.

ZestyAI today announced that Standard Casualty Company, a specialized property insurer serving manufactured homeowners, is expanding its partnership by adopting ZORRO Discover™, bringing agentic AI to regulatory and competitive intelligence.

Standard Casualty first partnered with ZestyAI in 2024, adopting Z-PROPERTY™, Z-FIRE™, and Z-HAIL™ for property-level underwriting. In 2025, the carrier expanded its use of Z-PROPERTY to include Roof Age and Wildfire Mitigation Prefill, applying them at the portfolio level to strengthen its view of risk across the book. 

With the addition of ZORRO Discover, the carrier is extending its use of AI across the full product lifecycle, from underwriting precision to rate strategy, competitive benchmarking, and state-by-state filing execution.

Historically, understanding competitor rate filings meant teams manually combing through thousands of pages of regulatory documentation across jurisdictions. ZORRO Discover automates that process, drawing on more than 2 million P&C rate and form filings to help regulatory, actuarial, and product teams spot rate trends, benchmark competitors, and get ahead of potential regulator objections.

For Standard Casualty, that means deeper competitive visibility, stronger pricing strategy, stronger submission readiness, and fewer delays from objection cycles.

Rick Smith, Underwriting Director at Standard Casualty, said:

"The ZestyAI platform has become core to how we underwrite and manage risk. ZORRO Discover was the natural next step, giving us structured visibility into how the market is moving and helping us strengthen our products and move through filing cycles more efficiently."

To support that kind of day-to-day decision making, ZORRO Discover goes beyond traditional research tools. It continuously analyzes new submissions and regulatory outcomes, so teams can track market shifts as they happen and adjust strategy proactively instead of reacting after the competition has already moved.

Attila Toth, Founder and CEO of ZestyAI, said:

‍"Insurance is quickly moving toward AI as core infrastructure. Standard Casualty has been putting that into practice across their business for years — and with ZORRO Discover, they're turning market moves and regulator signals into structured intelligence they act on in minutes instead of months."

ZestyAI today announced that NRDC (Natural Resources Defense Council), one of the nation's leading environmental advocacy organizations, is using ZORRO Discover™ to support its climate research and public policy advocacy efforts.

NRDC's FAIR Future Team uses ZORRO Discover to analyze insurance rate filings, track regulatory trends, and support climate-focused policy advocacy across U.S. insurance markets.

As climate-related risks reshape property insurance markets, policymakers and advocates face growing pressure to respond with evidence-based solutions. Yet the underlying data has historically been difficult to access—buried within hundreds of pages per filing, across dozens of insurers and multiple states.

ZORRO Discover addresses this challenge by using agentic AI to aggregate and structure more than 2 million P&C rate and form filings, representing over 200 million pages of regulatory documentation, into a unified, searchable system of decision intelligence.

NRDC will use the platform to support state-level advocacy and analyze rate and risk trends in key jurisdictions.

Alfonso Pating, Global Financial Regulations Specialist at NRDC, said:

“Insurance filings contain critical information about how risk is being priced and why—but extracting that information across dozens of companies and multiple states has traditionally required an enormous investment of time and effort."

NRDC's decision reflects a broader shift: regulatory data is no longer just a compliance artifact—it is a strategic asset for understanding how insurance markets function and evolve.

ZORRO Discover brings the same depth of regulatory intelligence used by insurers into the policy and research ecosystem shaping the industry.

Attila Toth, Founder and CEO of ZestyAI, said:

"Insurance filings contain the most detailed record of how risk is priced—but until now, they haven’t been accessible or usable at scale. That changes how insurance markets can be understood. We’re pleased to support NRDC’s work in this area."

ZORRO Discover continuously analyzes new regulatory submissions and outcomes across jurisdictions, giving users a current view of how insurance markets are changing. For NRDC’s FAIR Future Team, this means monitoring how insurers are responding to climate-related risk in real time—and to ground advocacy and media engagement in comprehensive, current data rather than fragmented or outdated sources.

The Zest: Key Takeaways

• Guy Carpenter's Kevin Van Leer on how property-level data is reshaping reinsurance placement and what's changing for catastrophe modelers.
• A broader shift is underway: carriers are elevating property data quality to an enterprise priority, sharpening the data that powers every decision across the insurance lifecycle — from underwriting and pricing to claims, portfolio management, and reinsurance.
• The bar for AI in insurance has shifted from "interesting" to "purpose-built": systems that are faster, more consistent, and more defensible than the manual processes they replace.

Few people have seen property analytics from as many angles as Kevin Van Leer. After studying atmospheric science at Purdue, he drove the release of the wildfire catastrophe model at RMS, a model whose appetite for granular property inputs ran ahead of what the data side could deliver at the time. He then spent seven years on the property analytics side at CAPE Analytics before joining Guy Carpenter as a Certified Catastrophe Risk Analyst. Today, he sits at the center of one of the most consequential shifts in reinsurance placement in a generation: how brokers, cedents, and reinsurers actually use property-level data to make capital decisions.

In a recent keynote conversation with the ZestyAI team, Kevin shared his view of where this market is going. The themes he covered are worth carrying forward.

Property data quality is now an enterprise imperative 

For decades, the industry has been moving from portfolio averages and class plans to property-level precision — and that shift is now reaching the part of the lifecycle where the most capital is at stake. Property-level data has long delivered value in underwriting and pricing, but the next leap is consistency: carriers building a single, trusted view of property risk and applying it the same way through underwriting, rating, claims, portfolio management, and reinsurance. The result is enterprise-grade data quality that holds up under reinsurer scrutiny. 

You can see the shift in how carriers go to market. The submission a cedent brings to reinsurers is a tight document covering financial highlights, key initiatives, exposure updates, and CAT loss updates, and data enrichment now sits among those headline sections rather than as a footnote.

What this unlocks is a virtuous lifecycle. Carriers sharpen their view of risk, make better risk selection decisions, capture appropriate rate, and underwrite with greater precision. When they bring that same view of risk to their reinsurer, both sides are aligned — giving cedents the clarity to make the right decisions on reinsurance structure, terms, and pricing. The view of risk gets sharper for everyone involved in the transaction. 

The CAT modeling gap is closing

Catastrophe models simulate peril at remarkable resolution, capturing wind fields, fire spread, flood inundation, and seismic intensity. They work best when those peril views are paired with equally granular property inputs, such as roof material captured at the address, building age that matches reality, defensible space measured rather than estimated, and exposures captured at the building rather than averaged across a ZIP code. For most of the industry's history, that level of property detail wasn't available at scale.

Much of this gap comes down to secondary modifiers — characteristics like roof material, roof age, roof condition, defensible space, and surrounding vegetation. These modifiers don’t come packaged with a CAT model, but they have a material impact on stochastic loss results. Without accurate, property-level secondary modifiers, even the most sophisticated model is making assumptions about the very inputs that drive its outputs. With them, carriers and reinsurers get a sharper, more defensible view of modeled losses.

As property-level data flows into reinsurance submissions, CAT modelers are working with the inputs their models were designed for, and reinsurers are pricing risk against a more accurate picture of what they're actually covering.

Purpose-built AI is the only kind that moves insurance forward 

Another theme that resonated: generic AI doesn’t move insurance forward. Purpose-built AI does. By “purpose-built,” Kevin means AI engineered for a specific insurance problem — replacing something expensive, slow, or inconsistent (manual inspections, piecemeal public-records pulls, or subjective desktop reviews) with something faster, more consistent, and uniformly applied to every property. 

The bar for adoption is simple: the AI’s output has to tie directly to claims outcomes. If an underwriter, actuary, or chief risk officer can see how a capability sharpens loss prediction, it earns its place in the workflow. If they can’t, no amount of polish makes it useful. 

What's next

There's a lot still ahead, including new markets to enter, new perils to model, and deeper integration into the systems where carriers and reinsurers actually make decisions.

Insurance is 700 years old, and the next decade is its most consequential yet. Property risk modeling sits at the center of that decade, combining technical and commercial work at the intersection of science and capital. 

Want to learn more about how ZestyAI helps carriers, brokers, and reinsurers optimize reinsurance decisions? Learn more about ZestyAI’s reinsurance solutions →