What I Learned at ITC: How GeoAI Is Rewriting the Rules of Wildfire Risk Modeling for Insurers

I just got back from InsureTech Connect Vegas, the world’s largest insurtech conference, where wildfire risk dominated nearly every conversation with modelers, underwriters, and data teams. But for me, this topic isn’t theoretical, it’s personal.

I live in Boulder, Colorado, where the Marshall Fire stopped just a mile and a half south of our home. Our CTO lost his house that day. The following year, our own homeowners premium jumped 50%, not because anything changed on our property, but because our insurer dropped a digital pin on our ZIP code and raised rates for everyone nearby.

That’s what happens when risk is priced by history instead of ground truth. It’s unfair to homeowners who invest in mitigation and frustrating for insurers forced to price uncertainty. At ITC, it became clear I’m not the only one thinking this way; the entire insurance industry is re-examining how we model and manage wildfire risk.

From black boxes to transparency

For decades, wildfire risk models were proprietary black boxes. Insurers relied on third-party vendors whose assumptions were rarely disclosed, even to regulators. But after a decade of escalating billion-dollar fire seasons and the mass withdrawal of insurers from high-risk markets, states are forcing a reckoning.

In California, regulators are rewriting the rules. Insurers must now disclose their wildfire risk models, explain how they work, and allow homeowners to see and appeal their individual risk scores. California is also building a public wildfire catastrophe model designed to benchmark private models used in rate filings.

This shift was triggered by necessity. A growing number of Californians live in ZIP codes where private insurers are scaling back or exiting the wildfire-exposed market, and premiums in high-risk ZIP codes have doubled or tripled since 2017. Transparency is meant to bring fairness back by tying premiums to actual property conditions rather than geography alone.

Colorado’s ground-truth approach

Here in Colorado, lawmakers are taking a similar but uniquely local approach. The Wildfire Risk Modeling Act, signed into law in May 2025 and taking effect July 1, 2026, requires insurers to disclose risk scores, mitigation discounts, and appeal processes directly to policyholders.

The law also compels companies to show how homeowner and community mitigation, like vegetation management and defensible-space improvements, actually change those scores. Together with new wildfire-resilient construction standards and large-scale utility mitigation plans, Colorado is becoming a testbed for the next generation of climate-resilient insurance.

How GeoAI changes the game

This is where GeoAI, or Geospatial AI, enters the picture. By combining satellite imagery, multispectral and LiDAR data, and machine learning, GeoAI allows insurers to analyze the built and natural environment with unprecedented detail.

Instead of relying on coarse fire-zone boundaries, models can now map roof materials, vegetation density, slope, and defensible space for each parcel. They can also detect change over time, such as new mitigation work or post-fire recovery, enabling dynamic underwriting rather than static, one-time assessments.

The result is a shift from reactive pricing to predictive understanding. Insurers gain more accurate loss ratios and regulatory compliance; homeowners gain transparency and the ability to lower premiums through measurable action.

The new frontier of risk modeling

Progressive MGAs, carriers, and data-driven startups are already showing how this future looks. They use AI-enhanced wildfire models to expand coverage in previously uninsurable areas and to reward mitigation rather than penalize proximity.

This new generation of modeling is not about replacing actuaries or engineers; it’s about equipping them with real-world context. When underwriters can see what a home is made of, how vegetation changes through the seasons, and whether a mitigation plan actually reduces risk, they can make smarter, more equitable decisions.

Why it matters

Wildfire modeling is shifting from opaque, backward-looking actuarial averages to explainable, physics-aware intelligence. That’s not just a technical upgrade; it’s a social and economic one. Transparent, data-rich models restore trust between insurers, regulators, and the communities they serve.

For homeowners, it means being judged by what’s real: the roof you replaced, the trees you trimmed, the defensible space you built, not by the worst fire miles away. For insurers, it’s the path back to sustainable underwriting in high-risk states. For regulators, it’s proof that climate adaptation and financial stability can coexist.

The takeaway

Wildfire modeling is shifting from black-box actuarial estimates to physics-aware intelligence, models that see, measure, and explain the world they price. The next era of insurance will belong to those who combine AI transparency with geospatial precision, pricing risk not by ZIP code, but by truth on the ground.

Originally published on LinkedIn