Cape Analytics for P&C Underwriting

In-Depth Review

Cape Analytics applies computer vision to aerial and satellite imagery to extract property-level data for insurance underwriting. The core use case: instead of sending an inspector to evaluate a roof, or relying on what an applicant self-reports on a homeowners application, a carrier queries Cape’s API and gets a roof condition classification, roof geometry measurements, and a set of hazard feature flags derived from overhead imagery.

What Cape Analytics Does for Insurance Professionals

The platform addresses a specific friction point in residential property underwriting: you cannot underwrite what you cannot see, and physically inspecting every roof in a book of business is not economically viable. Cape Analytics makes aerial imagery analysis available as an API call that returns structured data — roof condition score, pitch, area, vegetation proximity, pool and trampoline flags, solar panel detection — in seconds, at the point of quote.

This is distinct from aerial inspection services that produce a PDF report and require a human to review it. Cape’s output is machine-readable data designed to feed directly into underwriting workflows, pricing models, and policy management systems.

Key Capabilities

Roof condition scoring is the flagship feature. Cape classifies roofs as Good, Fair, or Poor based on visual indicators detectable from aerial imagery: missing or damaged shingles, granule loss, visible weathering, and surface irregularities. The score is updated as new imagery is acquired for an area — typically on a 12-to-24-month refresh cycle, though this varies significantly by geography.

Roof geometry extraction provides pitch, footprint area, shape complexity, and story count. These measurements have two direct applications: feeding replacement cost estimators with imagery-derived data rather than applicant-reported square footage, and providing CAT model inputs that reflect actual structure geometry rather than defaults.

Hazard feature detection covers the attributes that affect liability and peril-specific risk: vegetation within a configurable distance of the structure, pools, trampolines, and detached outbuildings. These are features that often go undisclosed on applications and that affect both pricing and eligibility decisions.

Solar panel detection identifies rooftop installations that affect replacement cost, roof access during a claim, and fire risk profile. This is increasingly relevant as residential solar penetration grows in certain territories and carriers find their policy data increasingly out of sync with the actual property.

Portfolio analytics lets carriers run their existing book against Cape’s data in batch mode — useful for renewal risk stratification (find policies where roof condition has deteriorated since last inspection), CAT exposure analysis (identify concentration of aging or damaged roof stock in storm-prone areas), and post-event claims pre-triage.

Pricing

Cape Analytics sells exclusively through enterprise contracts. Pricing is not publicly available. Deals are typically structured around API call volume (for new business enrichment) or portfolio size (for batch processing). Before entering a contract discussion, ask specifically about: minimum annual commitments, per-call pricing at your projected volume, imagery currency in your primary writing territories, and what happens when you query a property for which Cape has no imagery.

Honest Assessment

The value proposition is clear: automated roof condition data reduces underwriting decision time and catches undisclosed property conditions that affect risk. The questions worth stress-testing before committing are:

First, imagery currency. Cape’s data is only as useful as its underlying imagery, and in rural or lower-density markets that imagery may be two or three years old. Ask for a coverage map and representative sample dates for the geographies where you write the most business before assuming the data is actionable.

Second, score calibration. Cape’s Good/Fair/Poor classification is derived from image analysis, not a physical inspection. Carriers need to determine internally what Cape score threshold triggers a surcharge, a mandatory inspection, or a declination. That calibration requires running Cape scores against your own historical claims data to see how well the score predicts loss outcomes in your book — not just in aggregate industry statistics.

Third, error rates on non-standard roofs. The model performs best on standard asphalt shingle residential roofs. Complex roof types, flat commercial roofs, and heavily tree-canopied properties have higher classification uncertainty. Know your book’s mix before treating the score as a hard filter.

Who This Is Best For

Carriers writing standard residential homeowners at meaningful volume — roughly 20,000 or more policies — in well-imaged geographies are the natural fit. The per-call cost is justified when it replaces manual inspection costs or when it catches property condition issues that would otherwise result in adverse loss ratios. Regional carriers with dense concentrations in major metro areas and suburbs will find imagery currency more reliable than those writing in rural or mountainous territories.

One Thing to Test Before Committing

Request a sample data pull on 200 to 500 addresses from your current book — a mix of properties with known roof condition, ideally including some with confirmed claims in the last three years. Compare Cape’s condition scores against your internal records and any available inspection photos. Pay particular attention to the Fair-to-Poor boundary: if the model misclassifies a material number of claims-generating properties as Fair rather than Poor, your underwriting filter will be set on miscalibrated data. That test, on your own book, is more informative than any industry benchmark the vendor can provide.