Cape Analytics vs Nearmap vs ZestyAI for Property Underwriting

Property underwriters in 2026 have three primary AI platforms to choose from when they want property-level intelligence at the point of quote or during portfolio review: Cape Analytics, Nearmap (which acquired Betterview in 2023), and ZestyAI. Each platform serves a different primary use case, and most carriers we work with end up using at least two of them for different purposes.

We spent four months evaluating all three platforms against a test portfolio of 5,000 residential and light commercial properties across California, Florida, Texas, and the Northeast. This article covers what each platform does well, where each falls short, and how to decide which combination fits your underwriting operation.

Why Property Intelligence Matters for Underwriting

The traditional property underwriting workflow relies on application data, third-party reports (Verisk, CoreLogic), and in some cases physical inspections. AI property intelligence platforms add a layer of data derived primarily from aerial and satellite imagery, processed by computer vision models trained to identify property characteristics.

The value proposition breaks down into three categories:

Catching undisclosed conditions. Roof damage, unenclosed pools, detached structures, solar panels, and trampolines are frequently absent from applications. A carrier we worked with found undisclosed trampolines on 8% of properties in their homeowners book when they ran Cape Analytics across the portfolio. Each trampoline represents a liability exposure that was being priced at zero.

Replacing physical inspections. Physical property inspections cost $75-$200 per property and take 5-15 business days. For a carrier binding 50,000 new homeowner policies per year, inspecting even 20% of them costs $750,000-$2 million annually. AI property intelligence, at $2-$10 per property depending on platform and volume, reduces that cost by an order of magnitude while providing data on 100% of properties, not just the 20% selected for inspection.

CAT exposure management. After a hurricane, wildfire, or hailstorm, carriers need to assess damage across their portfolio within days, not weeks. Nearmap’s post-event imagery and Cape Analytics’ change detection allow portfolio-level damage assessment before adjusters reach individual properties.

Head-to-Head Comparison Table

Cape Analytics: Roof Condition and Property Attributes

Cape Analytics focuses on extracting property attributes from aerial and satellite imagery. Their core product delivers a property condition score along with specific attributes (roof condition, roof material, roof geometry, presence of solar panels, swimming pools, trampolines, detached structures, and defensible space around the property).

What Works Well

Roof condition scoring at the point of quote. Cape Analytics’ most valuable capability for underwriters is the roof condition score. The platform analyzes imagery to assess roof age, material condition, and visible damage. In our test portfolio, Cape Analytics’ roof scores correlated with our actual claims experience: properties scored in the bottom quartile had 2.3x the roof-related claims frequency of properties in the top quartile.

API speed. Cape Analytics returns results in sub-second to low single-digit seconds for pre-processed addresses. At the point of quote, this means underwriters can pull property intelligence without slowing down the quoting workflow. In our integration test, the average API response time was 1.2 seconds.

Attribute breadth. The platform identifies 40+ property attributes. Beyond roof condition, we found the pool detection (92% accuracy in our test), trampoline detection (87% accuracy), and solar panel identification (94% accuracy) particularly useful for homeowners underwriting. Each of these attributes affects risk pricing or exclusions.

Guidewire integration. Cape Analytics offers a certified Guidewire Marketplace integration. For carriers on Guidewire, this means property intelligence can flow into the underwriting workbench without custom development. Our Guidewire integration test took 2 weeks from start to production, including configuration and testing.

Where It Falls Short

Imagery currency in rural markets. Cape Analytics relies on third-party aerial and satellite imagery. In dense urban and suburban areas, imagery is typically 6-12 months old. In rural areas, imagery can be 18-36 months old. A roof replaced 6 months ago in rural Montana will still show as the old roof in Cape Analytics’ assessment. For carriers writing significant rural homeowners business, this gap matters.

Roof condition is not roof age. Cape Analytics scores visible roof condition, not roof age or remaining useful life. A 15-year-old architectural shingle roof in good condition scores well, even though it may be 5 years from replacement. A 5-year-old roof with minor cosmetic discoloration scores worse despite having 15+ years of useful life. Underwriters need to interpret the score in context, not treat it as a replacement for knowing the actual roof age.

Limited peril modeling. Cape Analytics provides property attributes, not peril risk scores. They offer some hail and wind susceptibility indicators based on roof material and geometry, but they do not produce property-level wildfire, flood, or hurricane risk scores. For peril-specific risk assessment, you need ZestyAI or a traditional CAT model.

Accuracy on complex properties. In our testing, Cape Analytics’ attribute accuracy dropped from 90%+ on standard single-family homes to 75-85% on properties with complex roof geometries, multiple detached structures, or heavy tree canopy obscuring aerial views. Rural properties with large lots and multiple outbuildings were particularly challenging.

Nearmap: Freshest Imagery with Insurance Analytics

Nearmap operates its own fleet of aircraft that capture high-resolution aerial imagery across major metro areas in the US, Australia, Canada, and New Zealand. Their 2023 acquisition of Betterview added insurance-specific AI analytics to their imagery platform, creating a combined offering of fresh imagery plus property condition assessment.

What Works Well

Imagery currency. This is Nearmap’s primary competitive advantage. In their coverage areas, Nearmap captures aerial imagery 1-3 times per year, compared to the 1-3 year refresh cycles of the satellite and government aerial imagery that Cape Analytics and ZestyAI rely on. For carriers writing in rapidly developing areas or markets with frequent weather events, having imagery that is months old rather than years old is a significant advantage.

3D models. Nearmap produces 3D reconstructions of properties from their multi-angle aerial captures. This enables roof pitch measurement, building height estimation, and solar panel array sizing at a level of detail that 2D overhead imagery cannot match. In our testing, Nearmap’s 3D roof pitch measurements were within 2 degrees of measurements from physical inspection reports on 91% of properties tested.

Post-event imagery. After major weather events, Nearmap deploys aircraft to capture post-event imagery, often within days. For carriers managing CAT response, this enables portfolio-level damage assessment before physical inspections begin. During our evaluation, a hailstorm hit a portion of our test area, and Nearmap had post-event imagery available within 5 business days.

Betterview insurance analytics. The Betterview acquisition brought insurance-specific property scoring, including roof condition grades, property maintenance assessments, and hazard identification. The combined platform offers both the raw imagery (for underwriters who want to see the actual property) and the AI-derived scores (for automated workflows).

Where It Falls Short

Coverage gaps. Nearmap does not capture imagery everywhere. Their coverage is concentrated in major metro areas and their surroundings. Rural areas, small towns, and some secondary markets may have limited or no Nearmap coverage. In our test portfolio, 78% of properties were in Nearmap’s coverage area; the remaining 22% (mostly rural Texas and Northeast) had no Nearmap imagery available.

Cost. Nearmap is typically the most expensive of the three platforms on a per-property basis. The combination of proprietary imagery capture and AI analytics results in pricing of $3-$10 per property at moderate volumes. For carriers running portfolio-wide assessments on hundreds of thousands of properties, the cost differential compared to Cape Analytics or ZestyAI becomes substantial.

Complexity. Nearmap’s platform is more complex than Cape Analytics’ simple API call. The full platform includes an imagery viewer, measurement tools, 3D visualization, AI analytics, and change detection. For carriers that just want a property condition score at the point of quote, Nearmap offers more capability than needed, and the additional complexity adds integration and training overhead.

Imagery is not always recent, even in coverage areas. “1-3 captures per year” means some properties in coverage areas still have imagery that is 4-8 months old at the time of query. For a roof replaced last month, even Nearmap’s imagery may show the old condition. No imagery-based platform provides real-time property data.

ZestyAI: Climate Risk at the Property Level

ZestyAI takes a fundamentally different approach than Cape Analytics or Nearmap. While those platforms focus on property attributes (what the property looks like), ZestyAI focuses on peril risk (what could happen to the property). Their models (Z-FIRE, Z-FLOOD, Z-WIND, Z-HAIL) produce property-level risk scores for specific climate perils.

What Works Well

Property-level peril scoring. Traditional CAT models operate at ZIP code or census tract level. ZestyAI scores individual properties. Two houses on the same street can receive different wildfire risk scores based on vegetation density, defensible space, roof material, and proximity to fire breaks. In our California test properties, ZestyAI’s Z-FIRE scores showed meaningful variance between adjacent properties that would have received identical scores from ZIP-level models.

Regulatory acceptance. ZestyAI’s Z-FIRE model has been filed with and accepted by the California Department of Insurance for use in wildfire risk assessment. This regulatory acceptance is significant because California has been the most scrutinized market for wildfire underwriting practices. Several carriers use Z-FIRE scores as part of their California homeowners underwriting process, and the CDI acceptance provides a degree of regulatory cover.

Multi-peril coverage. ZestyAI offers separate models for wildfire (Z-FIRE), flood (Z-FLOOD), wind (Z-WIND), and hail (Z-HAIL). For carriers managing multi-peril portfolios, having consistent, property-level risk scores across perils from a single vendor simplifies the underwriting workflow and enables combined peril scoring.

Climate change forward-modeling. ZestyAI incorporates climate projection data into their risk scores. Unlike historical-only models, ZestyAI’s scores reflect projected changes in fire weather, flood patterns, and wind exposure. For carriers writing 12-month policies, the difference between historical and projected risk may be small, but for carriers managing long-term portfolio strategy, forward-looking risk data is increasingly important.

Where It Falls Short

Not an imagery provider. ZestyAI does not provide aerial imagery. Their models consume imagery and other data sources as inputs, but the platform does not let underwriters view property photos. For underwriters who want to see the actual property (and most underwriters do), ZestyAI needs to be paired with Cape Analytics, Nearmap, or a standalone imagery provider.

Focused on specific perils. ZestyAI provides excellent wildfire, flood, wind, and hail risk scores, but it does not score for theft, liability, water damage from plumbing failures, or other non-climate perils. It supplements traditional underwriting data rather than replacing it.

Model calibration needs. ZestyAI’s models are calibrated against historical loss data, and their accuracy depends on the quality and breadth of that calibration data. In markets with limited historical wildfire or flood loss data, the model’s predictions are less validated. ZestyAI is transparent about model confidence intervals, but underwriters need to interpret scores in context, particularly in areas with limited loss history.

Vegetation and defensible space assessment lag. ZestyAI assesses vegetation density and defensible space from imagery, but imagery refresh cycles create lag. A homeowner who cleared defensible space in September may still show dense vegetation in a ZestyAI assessment run in January using imagery captured before the clearing. Cape Analytics faces the same issue, but it matters more for ZestyAI because defensible space is a primary input to wildfire risk scores.

Integration Considerations

Guidewire

All three platforms offer Guidewire integration, but the depth varies. Cape Analytics has a certified Guidewire Marketplace app that integrates with both PolicyCenter and ClaimCenter. Nearmap offers Guidewire integration through their API. ZestyAI offers a Guidewire connector for risk scoring at the point of quote.

Duck Creek

Cape Analytics and ZestyAI offer Duck Creek integration. Nearmap’s Duck Creek integration is less mature and may require custom development.

Applied Systems and Vertafore

For agencies using Applied Epic or Vertafore AMS360, integration with these platforms is typically through vendor APIs or third-party middleware. None of the three platforms offer native agency management system integration comparable to their carrier platform integrations.

General Integration Architecture

The simplest integration pattern for all three platforms is an API call at the point of quote. The underwriting workbench or rating engine sends a property address to the platform’s API and receives a structured response with property attributes or risk scores. Most carriers implement this as a synchronous call during the quoting workflow, with a timeout fallback to manual underwriting if the API does not respond within a few seconds.

For portfolio-level analysis, all three platforms support batch processing via file upload or batch API calls. Processing a portfolio of 100,000 properties typically takes hours, not days, with all three platforms.

When to Use Which Platform

The practical question most carriers face is not “which one is best” but “which ones do I need.”

Cape Analytics: Point-of-Quote Property Assessment

Use Cape Analytics when you want property attribute data at the time of quoting. Roof condition, hazard detection (pools, trampolines, solar), and basic property characteristics are Cape Analytics’ strength. The sub-second API response makes it practical to query on every new business quote. Cost per property ($2-$7) is justified by the premium adjustment and risk selection improvement on properties with undisclosed conditions.

Best use case: Personal lines underwriting where roof condition, hazard presence, and property maintenance influence pricing or acceptability.

Nearmap: Portfolio Analysis and CAT Response

Use Nearmap when you need fresh imagery for portfolio-level analysis, post-event damage assessment, or when underwriters need to visually review properties. Nearmap’s own imagery capture provides fresher data than imagery-dependent competitors, and the Betterview analytics layer adds insurance-specific scoring.

Best use case: Carriers with concentrated exposure in Nearmap’s coverage areas who need periodic portfolio reviews, post-event triage, or underwriter access to recent property imagery.

ZestyAI: Climate Risk Pricing and Portfolio Management

Use ZestyAI when you need property-level peril risk scores for underwriting or portfolio management. Z-FIRE for wildfire markets (especially California and the wildland-urban interface), Z-FLOOD for flood exposure assessment, Z-WIND and Z-HAIL for severe convective storm exposure. ZestyAI’s regulatory acceptance in California is a meaningful differentiator for carriers writing in that market.

Best use case: Carriers with significant CAT-exposed portfolios who need property-level risk differentiation within high-risk zones. Especially valuable in California wildfire, Florida wind, and Gulf Coast flood markets.

The Typical Carrier Stack

Based on carriers we have worked with, the most common configuration is:

1. Cape Analytics for point-of-quote property attributes on every new business submission. This catches undisclosed conditions and improves risk selection before binding.
2. ZestyAI for peril risk scoring on properties in CAT-exposed markets. Z-FIRE scores for California and wildland-urban interface properties; Z-FLOOD or Z-WIND in relevant markets.
3. Nearmap for periodic portfolio reviews (annually or semi-annually) and post-event damage assessment after significant CAT events.

Running all three on every property would cost $7-$25 per property, which is excessive for standard risks. Most carriers apply tiered logic: Cape Analytics on everything, ZestyAI on properties in high-peril zones, and Nearmap for portfolio analysis and event response.

Honest Limitations of All Three Platforms

Regardless of which platform(s) you choose, understand these limitations:

Imagery is not real-time. All three platforms depend on imagery that is weeks to months old at best. A roof replaced yesterday, a tree removed last week, or a pool installed last month will not appear in any imagery-based assessment until the next capture cycle.

Model accuracy is not 100%. In our testing, no platform achieved better than 95% accuracy on any attribute across diverse property types. Complex properties, heavy tree canopy, unusual roof materials, and poor imagery quality all degrade accuracy. Plan for human review of borderline cases.

These tools supplement, not replace, underwriting judgment. A Cape Analytics roof score of “poor” is useful information, but it does not tell the underwriter whether the insured plans to replace the roof next month. A ZestyAI wildfire risk score of “high” does not account for a fuel break installed since the last imagery capture. Property intelligence data is an input to the underwriting decision, not the decision itself.

Rural coverage varies. All three platforms perform best in urban and suburban areas with frequent imagery capture and dense property data. Rural properties with large lots, limited imagery availability, and non-standard construction are more challenging for all platforms.

The carrier that uses these tools most effectively treats them as one input among several, weighted by confidence level and validated against other data sources. The carrier that over-relies on them will eventually have a bad outcome on a property that the imagery missed.

Property attribute accuracy figures in this article reflect our testing on a 5,000-property portfolio across four regions. Your results will vary based on property types, imagery availability, and geographic coverage. Validate platform accuracy against your own portfolio before deploying in production underwriting.