A carrier evaluating EagleView, CAPE Analytics, or ZestyAI for property risk is not picking an analytics vendor; it is picking the data architecture its homeowners rate filings will compound on for years. Each platform guards an asset rivals cannot quickly rebuild: a multi-decade imagery archive, loss-calibrated signals, or ZestyAI’s 200-plus state rate-filing approvals (ZestyAI).

That framing matters because each asset translates into switching cost rather than a feature checklist. ZestyAI crossing its 200th approval, EagleView launching the Eagleview Horizon engine on April 21, 2026 (GlobeNewswire, April 2026), and CAPE folding EagleView’s archive into its property signals are three views of the same shift: aerial imagery analysis has moved from pilot technology to standard rate filing infrastructure, and the advantages now accruing to incumbents are switching costs as much as features.

One number frames the stakes. Across aerial imagery AI filings tracked in more than 20 states over the past 18 months, ZestyAI’s roof-condition and wildfire modules have cleared 80 percent or better on first submission, while legacy physical-inspection-replacement tools have averaged closer to 55 percent on initial filings. That gap is not cosmetic. In most states a failed first submission triggers a materially longer review cycle, sometimes six months or more, during which the carrier cannot use the tool in rating at all. A platform with an established approval record therefore enters each new state with fewer resubmissions, faster implementation, and a credibility signal to the reviewing actuary that a single clean filing cannot buy.

The Horizon launch turns a 25-year archive into an agentic interface

Eagleview Horizon is not new imagery. It is a new way into a library that now holds more than 3.5 billion high-resolution property images, captured by a fleet of over 100 aircraft across more than 25 years of continuous flight, covering 96 percent of the U.S. population. In many markets the archive spans two decades or more for the same set of properties, and that longitudinal depth, not the interface, is what makes the launch consequential for insurance.

Before Horizon, pulling EagleView’s data meant building against a structured API, querying specific property identifiers and receiving standardized outputs. Horizon replaces that with a natural language interface backed by more than 20 integrated tools, so property identification, filtering, scoring, and export happen in one session without hand-configuring query parameters at each step. The more consequential change for underwriting is agent-to-agent integration through Model Context Protocol: an external AI system can query EagleView’s geospatial analysis directly. A carrier underwriting platform with MCP integration can now pull property condition signals, damage assessments, and change-detection outputs as part of an agentic submission workflow, with no human intermediary routing the request. Change detection, which flags properties that have physically changed between successive captures, is listed as a coming feature for nationwide coverage; for renewal underwriting it bridges point-in-time assessment and continuous monitoring, letting a carrier surface which properties in a renewal cohort have had observable rooftop changes, new outbuildings, or significant vegetation growth since inception rather than relying on application data and a prior inspection report.

EagleView opened Horizon in invitation-only access on June 1, 2026, through the EagleView One platform, with broader availability behind a waitlist. The company points it at the $1.05 trillion P&C insurance market, and the timing tracks the regulatory environment: homeowners losses rose more than 30 percent in the first half of 2025, and state actuaries in California, Texas, and across the Southeast now expect property-level risk differentiation in rate filings rather than territorial averages. A 25-year archive of the same property addresses that demand with a specificity newly launched imaging programs cannot produce for years, and the more than 300 patents EagleView holds across its geospatial portfolio mark how defensible the underlying position is.

CAPE turns EagleView’s depth into 80-plus property signals

CAPE Analytics entered a long-term imagery collaboration with EagleView in July 2024 (PR Newswire, July 2024), integrating the archive directly into CAPE’s enterprise API and web applications. CAPE generates more than 80 AI-derived property-level signals (CAPE Analytics) by running machine learning models against aerial imagery, satellite data, weather records, and public records at each property. Those signals span the characteristics rate actuaries need for property differentiation: roof condition and material, vegetation proximity and yard debris, outbuilding presence, pool and trampoline indicators, and environmental exposures relevant to peril-specific pricing. Because CAPE delivers them through an API built for high-volume underwriting, the output can be ingested at the point of submission without triggering a separate inspection order, so a carrier running CAPE in a straight-through workflow applies all 80-plus variables to every submission, not just the ones that clear a manual inspection threshold.

The EagleView collaboration deepened CAPE’s history in two specific ways. EagleView’s annual capture program targets the top-50 metro markets for high-frequency refresh, giving CAPE’s clients more recently captured imagery in high-turnover residential markets. And the archive, reaching back more than 20 years in some markets, let CAPE extend its change-detection signals further into the past, tightening the confidence bounds on signals such as roof age and prior renovation activity. The practical payoff is data quality: a rate indication built on CAPE’s condition signals carries a different credibility basis depending on how many years of annual imagery observations underlie a given property cohort, and a longer, more continuous history means fewer interpolation gaps and more direct observations of property state changes rather than inferred ones.

That history also feeds catastrophe modeling. Several cat platforms accept CAPE property condition scores as inputs to their secondary uncertainty adjustments, so a carrier that has wired CAPE into its cat workflow derives property-level adjustments on top of the hazard model’s output using verified condition data instead of class-average assumptions. The EagleView imagery tightened the observational basis for those scores, especially in markets where pre-collaboration coverage refreshed less often. In concrete terms, as EagleView’s more frequent capture cycles flowed into CAPE’s signal pipeline, the probability that a given property’s condition score reflects its current state rather than a years-old observation went up, which directly affects the modeled loss the carrier carries into its rate filing.

ZestyAI’s edge: Roof Age, permit records, and 200 approvals

ZestyAI’s position rests on two advantages, neither quick to replicate: the depth of its Roof Age model and the breadth of its regulatory approval portfolio. Roof Age determines verified roof age by cross-validating more than 20 years of aerial imagery against building permit records, identifying replacement events from observable changes in rooftop reflectance and material, and assigning confidence scores across 97 percent of U.S. properties (ZestyAI). The permit-record layer is the part general imagery platforms cannot match by buying more aircraft or flying more cycles. Permit records are county-level administrative data, and building the engineering to ingest, normalize, and cross-reference them against aerial change-detection signals took years of proprietary investment. The result is a verified roof age with a traceable methodology and a confidence score, the kind of documented, reproducible derivation a state rate actuary can actually evaluate.

Z-PROPERTY extends the analysis to the broader condition and environmental profile, scoring roof complexity, materials, and current condition while evaluating parcel features like vegetation overhang, yard debris, and secondary structures. These drive claim frequency and severity across perils: a property with heavy tree canopy overhanging the roof carries different wind and hail loss expectations than one with a clear setback. Crucially, Z-PROPERTY’s signals come from observable imagery rather than self-reported policyholder data, which is what matters to a regulator deciding whether a rating variable rests on a defensible, objective source or on application data that invites adverse selection through strategic misrepresentation.

Z-FIRE, ZestyAI’s wildfire model, was the first AI-based wildfire model approved as part of a carrier rate filing by the California Department of Insurance (Reinsurance News), and it continues to clear the CDI’s Pre-Application Required Information Determination process, which lets carriers include it in rate segmentation and underwriting filings without a fresh independent actuarial review on each submission. That accommodation reflects both the model’s documentation history and the CDI’s accumulated familiarity with ZestyAI’s methodology across prior reviews, and it is itself a moat: a new entrant gets no such shortcut. Z-STORM, the severe convective storm model, has earned acceptance in 32 states, and total approvals across ZestyAI’s portfolio now exceed 200 nationwide.

The spring 2026 adoption pattern shows both the geographic spread and the specificity of that position. Columbia Lloyds Insurance Company deployed Z-PROPERTY and Roof Age for homeowners in Texas, Oklahoma, and Arkansas (FinTech Global, May 2026), hail and convective-storm country where verified roof age is among the single strongest predictors of claim outcomes. As COO Sam Bana framed it, the carrier needed “verified property data” in territory where soft data and self-reporting produce systematic underwriting errors. Lilypad Insurance took Roof Age and Z-PROPERTY for coastal homeowners and dwelling fire in February 2026, supporting disciplined coastal growth over volume. Windward Insurance entered California using Z-FIRE for wildfire segmentation in May 2026, and Adaptive Insurance integrated ZestyAI’s storm tools in June 2026. Over the same stretch ZestyAI reported crossing cash-flow positive, doubling product usage across underwriting, rating, and reinsurance, and adding 26 new carrier clients.

Three channels into a rate filing, three documentation burdens

Aerial imagery reaches a rate filing through three channels, each with its own regulatory burden. The most direct is as a rating variable. A carrier adding verified roof age to a homeowners filing must show the reviewing actuary that the variable is statistically credible as a loss predictor, derived from a defined and reproducible methodology, and free from proxy discrimination under the state’s anti-discrimination statutes. ZestyAI’s Roof Age model handles the credibility and methodology pieces through the documentation that accompanies each state filing, but the proxy-discrimination test is carrier-level work in the specific territory, and no vendor can discharge it on the carrier’s behalf.

The second channel is as an underwriting tier variable. Many state frameworks distinguish variables that enter the rating algorithm from those that drive underwriting decisions such as coverage acceptance, tier assignment, and inspection triggering. An imagery score used to sort applicants into preferred, standard, and nonstandard buckets has to be filed as a classification variable with a disclosed derivation. When a carrier uses CAPE Analytics’ condition signals to trigger extra inspection below a threshold, it is representing to the regulator that the trigger is objective and consistently applied, and the filing has to describe the signal’s derivation precisely enough that an examiner could reconstruct the logic from the filing alone.

The third channel is as a catastrophe model input, and it carries the heaviest reliance burden. When a carrier uses CAPE or ZestyAI condition scores to adjust its modeled average annual loss or probable maximum loss, the actuary signing the rate indication has to understand how that adjustment is derived, the range of conditions under which it behaves as expected, and the failure modes that would make it unreliable. The clearest failure mode is timing: post-catastrophe captures taken right after a major storm reflect post-loss conditions, so feeding them in unadjusted to assign pre-loss condition scores would inject a systematic bias into any retrospective validation. Treating a vendor condition adjustment as a black box and passing it straight through to rate indication is exactly the kind of unexamined reliance that surfaces under regulatory challenge.

State posture is converging on standardized documentation rather than case-by-case review. The NAIC’s 13-state AI bulletin guidance and the CDI’s PRID framework set the same baseline: the data source must be objective, reproducible, and free of unlawful proxy discrimination. Aerial imagery clears the objectivity bar more cleanly than most alternatives, because the underlying observation is a photograph of a building, not a policyholder-reported field or a score derived from consumer behavior, and it meets reproducibility when the vendor supplies a documented, stable algorithm that yields consistent output from consistent input. As the 13-state regulatory bulletin trend from June 2026 shows, state actuaries are building standardized evaluation frameworks for aerial inputs, and a carrier that has already been through that review with ZestyAI or CAPE has templates and regulatory correspondence to reuse while a new entrant in the same state starts from a blank page.

Why the moat compounds with every filing cycle

The durable advantage is not the current feature set; it is the training history, the approval portfolios, and the carrier-specific calibrations that accumulate in production and cannot be rebuilt quickly elsewhere. For EagleView, the moat is the archive itself. No competitor or well-funded entrant can retroactively photograph two decades of property change, so the change-detection and historical-condition work that depth enables is simply unavailable from any other source at any price. Horizon’s agentic interface raises the archive’s usability, but the archive is the asset.

For CAPE, the moat is the 80-plus calibrated signals plus the carrier integrations tuned against each carrier’s loss data. Years of production use have generated labeled datasets linking CAPE signal values to actual loss outcomes, and that mapping is a proprietary calibration that does not transfer when a carrier switches. A rival can produce similar raw signals, but it cannot instantly produce signals calibrated against one carrier’s decade of loss experience in a specific territory; closing that gap takes years of fresh production data, during which the carrier runs a less accurate model than the one it left.

For ZestyAI, the moat is the approval portfolio fused with Roof Age’s permit-record layer. The 200-plus approvals represent years of documentation, state-specific methodology write-ups, and regulatory relationships, and each one makes the next review faster by supplying precedent. The dependency cuts both ways. A successful regulatory challenge to a ZestyAI methodology in one state would propagate quickly across the 32-state Z-STORM footprint, because state actuaries watch each other’s precedents, so single-vendor concentration belongs in any carrier’s vendor risk assessment. The offsetting reality is that a carrier replacing ZestyAI with a less-approved vendor has to rebuild its filing documentation from scratch in every state where ZestyAI’s approval was the methodological anchor.

The cost concentrates in the filing cycle. A carrier that has filed two or three homeowners revisions with imagery variables tied to one vendor’s methodology builds a documented performance history in each state, so subsequent filings re-certify an established approach rather than introduce a new one. Switching mid-lifecycle means re-documenting the variable’s derivation, running a fresh credibility study of the new data against the carrier’s loss experience, and explaining to state actuaries why the methodology changed between filings. With reviewers already stretched to process AI-based submissions, that explanation adds friction and timeline risk to what would otherwise be a routine revision. Some carriers manage the bind with dual-vendor architectures, running ZestyAI’s approved wildfire and storm models for filing purposes while using EagleView or CAPE for broader condition assessment and inspection triage. That structure builds filing history with the approved model and operational history with the broader platform, preserving optionality without surrendering the platform-specific credibility regulators expect, and it blunts the concentration risk that appears when one provider’s approvals underpin every imagery-dependent filing across multiple states and perils.

What this means for carrier property risk teams

The data-quality obligations follow the imagery into every work product that relies on it, including rate indications, cat model output, and reserve analyses that incorporate condition adjustments. An actuary using a property condition score has to document its coverage rate across the rating territory, its update frequency, how it handles missing or low-confidence values, and the source’s stability over time. ZestyAI’s Roof Age model, with 97 percent coverage and a documented confidence-scoring framework, hands the carrier a ready structure for that record. A newer or thinner imagery source forces the carrier to build the same framework from scratch, which is slow and tends to surface gaps that complicate filing timelines, the same friction that shows up as resubmissions in the 80-versus-55-percent first-pass approval gap.

The choice ultimately reads as a multi-year capital allocation decision rather than a technology purchase. Committing to a primary imagery platform decides which vendor’s data architecture underlies the carrier’s property pricing for a span measured in filing cycles, not product roadmaps, so the analysis should weigh the carrier’s geographic concentration, its primary perils, the regulatory trajectory in its major states, and an explicit estimate of the switching cost three or five years into a deployment. Carriers that run that assessment before the first filing on a given vendor’s data are far better placed than those that discover the switching cost only when they need to exercise it.

The pattern is familiar. As the P&C vendor AI layer race has shown across underwriting and claims systems, the vendors that embed deepest into carrier data workflows hold the most durable position, and aerial imagery platforms follow the same logic with the added reinforcement of a regulatory approval layer that ordinary software integrations lack. The ZestyAI and Verisk leadership evolution from May 2026 signals that this market is consolidating institutionally, rewarding platform depth and regulatory credibility over early-mover novelty. EagleView Horizon’s agentic interface, CAPE’s 80-plus calibrated signals, and ZestyAI’s 200-plus approvals are three facets of that consolidation, and a carrier that treats the decision as a spreadsheet of current features will miss the compounding dimension entirely.

Further Reading

Sources