Geospatial data reshapes underwriting

Property insurance used to smooth the world into broad zones. A ZIP code stood in for a neighborhood. A census tract stood in for a block. Sometimes a county stood in for an entire fire corridor. That was tolerable when underwriting was built around coarse maps, slow inspections, and public records that changed by the year. It looks much less defensible now that insurers can buy rooftop-level geocodes, building footprints, aerial imagery, and hazard layers that describe one parcel at a time. Recent industry research and insurer deployments all point the same way: carriers are updating risk profiles where local detail changes the answer, because local detail often changes the answer a lot. (milliman.com) The important shift is not just better pictures. It is better resolution. Milliman’s recent work for the actuarial societies argues that structure footprints derived from high-resolution imagery can improve risk determination in flood and wildfire settings because the precise location and shape of a building matter more than the center point of a parcel. The Society of Actuaries framed the same idea more bluntly: geospatial precision can validate prefilled underwriting data, identify ground features remotely, and reduce the need for site inspections. In property lines exposed to wildfire, flood, or roof-related losses, that turns underwriting from a regional guess into a property-level measurement problem. (casact.org) That matters most in markets where insurers have been pulling back. In May 2025, the Insurance Information Institute published a California wildfire study built around Guidewire HazardHub scoring and parcel-level analysis. Its finding was simple and disruptive: even inside communities broadly labeled high risk, there are lower-risk properties that older, blunter methods miss. Google Cloud made the same commercial case a few months later, arguing that geospatial AI lets carriers move beyond regional models and identify insurable risks in climate-exposed areas that competitors have abandoned. The point is not that climate risk has been overblown. The point is that risk is lumpy, and insurers now have tools to see the lumps. (iii.org) Once underwriting becomes that granular, insurers start shopping for different evidence. FEMA’s National Flood Hazard Layer is still foundational for flood decisions, but it is no longer enough on its own for many carriers. They are layering it with current aerial imagery, building-level geocodes, roof condition signals, vegetation proximity, parcel characteristics, and vendor-generated property intelligence. Zurich North America said in October 2025 that its middle-market underwriters were using Nearmap’s Betterview platform to pull in updated imagery, roof condition scores, deferred-maintenance signals, and other indicators that even physical inspections may miss or capture less consistently. That is a different underwriting workflow from asking for a loss run and sending an inspector. It is a workflow built around continuous external observation. (fema.gov) The new workflow is powerful enough that regulators have started policing it. Massachusetts told insurers in Bulletin 2025-02 that aerial imagery can support underwriting action, but not when the image is ambiguous, stale, or merely cosmetic. If imagery does not provide clear evidence of meaningful degradation, the state expects further review and may expect a physical inspection. West Virginia and New Hampshire issued similar guidance after complaints tied to homeowners coverage decisions based on aerial images alone. That tells you how fast the practice spread. The fight is no longer about whether insurers can use geospatial evidence. It is about how much they can trust it before a person has to knock on the door. (mass.gov) Behind that fight is a quieter commercial fact. The external data market around underwriting is being rebuilt to serve this demand. NICB’s Geospatial Insurance Consortium now pitches access to one of the industry’s largest aerial imagery libraries. Vendors like Ecopia, Vexcel, and Nearmap sell continuously refreshed building footprints, rooftop geocodes, and imagery-derived property attributes because insurers increasingly want evidence that is current, parcel-level, and machine-readable. Underwriting in new markets used to begin with territory manuals and broad hazard maps. More often now, it begins with a footprint on a roof. (nicb.org)