Leveling Up Property Risk Modeling: Reflections from the ZestyAI 2026 Offsite

‍

Last week, ZestyAI gathered in Cabo San Lucas for our 2026 company offsite under the theme Level Up. We were joined by Kevin Van Leer of Guy Carpenter — one of the most respected voices in catastrophe analytics and reinsurance — for a keynote conversation on where property risk modeling is headed.

A few themes we're carrying into our work.

Property risk modeling is entering its enterprise era

For years, third-party property data was used across pockets across the insurance lifecycle  — valuable for individual underwriting decisions, but rarely trusted at portfolio scale. That era is ending. Carriers are recognizing the need to enhance enterprise data quality: building a trusted, consistent view of property risk governed centrally and applied consistently across underwriting, rating, claims, portfolio management, and capital strategy.

When a consistent view of risk flows through every stage of the insurance lifecycle, a virtuous cycle takes hold. Carriers select better risks. They price those risks more accurately. They see the full picture of their exposure. The compounding effect shows up in combined ratios, in claims outcomes, in capital conversations — and, when it matters, in the defensibility of every decision they've made.

The "purpose-built" test for AI

One theme from Kevin's keynote landed especially hard with the room: AI for its own sake doesn't move the industry forward. What does is purpose-built tooling that replaces something expensive, slow, or inconsistent — manual inspections, piecemeal public-records pulls, subjective desktop reviews — with something faster, efficient, and uniformly applied to every property.

That's a useful test for any AI capability landing in front of an underwriter, an actuary, or a chief risk officer right now: Is this a better way to do something we already do? Or is it a demo? Solutions that pass the first test get adopted. The rest get politely ignored.

A Year of Momentum

The week capped a year of meaningful momentum. Over the past 12 months, ZestyAI:

• Turned cash flow positive while more than doubling product usage across underwriting, rating, and reinsurance workflows
• Welcomed 26 new clients, including Applied Underwriters, California Casualty, Lemonade, and Marsh
• Expanded commercial relationships with 12 existing enterprise customers through new agreements, including Berkshire Hathaway, the California FAIR Plan, and CSAA
• Surpassed 200 regulatory approvals nationwide, spanning multiple products and perils
• Launched Z-SPARK™, bringing modern fire science and property-level intelligence to how building materials, maintenance, surrounding structures, local fire response, and climate drive ignition and spread
• Introduced ZORRO Discover™, our agentic AI platform embedded into underwriting, portfolio management, and competitive intelligence workflows

That's where we're starting from. Here's where we're going next.

What "Level Up" means for ZestyAI

For us, Level Up is a commitment to meet customers where the market is going — not where it has been. In practice that means:

• Investing in continuous innovation — new products, new categories, new ways of solving problems carriers haven't yet found a way to solve 
• Helping carriers cover the losses that hit them hardest, with continued peril-science investment across the exposures driving today's largest claims and reinsurance decisions 
• Scaling ZORRO™ into a full agentic orchestration platform, embedding AI deeper into underwriting, portfolio management, and competitive intelligence workflows
• Building products that connect risk to capital, designed for the workflows where exposure data shapes capital and capacity decisions
• Expanding Regulatory & Government Affairs — building on 200+ approvals to be a strategic partner for carriers navigating an increasingly complex regulatory landscape, not simply a model provider