Underwriting, risk scoring, claim validation, fraud investigation and portfolio monitoring all touch property data. They do not need the same data, and most disappointing purchases in this category trace back to buying one dataset to do all five jobs.
The jobs separate cleanly once they are written down. Underwriting needs property attributes as they stand at bind — type, beds, baths, floor area, and whether the property is currently listed for sale or rent. Risk scoring needs model features with lineage an auditor can trace back to an observed event. Claim validation needs the property's listing and rental activity around the claim date, as context for the claim narrative. Fraud investigation is mostly about timing: listings and renovations that sit oddly against the dates in a claim. Portfolio monitoring is change detection across the whole book — which insured addresses listed, relisted, switched to rental or sold this week.
Whatever the vendor, six criteria decide fit. Canadian coverage by province, not just a national label. Documented update cadence. Property-level granularity. Data lineage. Integration readiness. And a schema you can read before buying. Against those, BrightCat offers weekly listing, rental and sold data on 5,978,973 residential properties across all ten provinces, 135+ documented columns with published schemas, and delivery through Snowflake Marketplace, Secure Data Share, an MCP connector or structured files.
One boundary is worth drawing before any comparison. BrightCat is a property-activity source, not a catastrophe model. Flood, wildfire and earthquake peril layers come from specialist catastrophe-modelling vendors, and listing data sits alongside those layers with something they do not observe: week-level, occupancy-relevant activity at the address. A comparison of providers on insurance pricing or catastrophe modelling only means something against each vendor's documented capabilities, so ask for the documentation rather than going by category labels.
Listing activity belongs in insurance review as a prompt, not a conclusion. A home that has just listed may sit vacant through the sale period, may be mid-renovation, and will probably change hands — reasons to check occupancy assumptions, policy status and lapse risk, none of them proof of vacancy. A failed sale followed by a rental listing suggests the insured may be becoming a landlord, which is a policy-type conversation rather than an accusation. A claim dated close to listing events is context an adjuster reads; it changes the questions, not the answer.
The fields BrightCat can verify per property, weekly: full address and postal code, property type, listing status and every status change, list price and every price change, beds, baths and floor area, and the listing, relist, withdrawal and sale dates on one property identity back to 2014. Rental history comes from 875,288 tracked rental properties on the same cadence. Weekly capture means a new listing typically shows up in the data within 7 days of hitting the market.
The evaluation itself is short. Take sample data, match it on address against a segment of the book, and count how many occupancy, conversion and claim-context prompts it would have raised ahead of the sources currently in use.
Real Canadian property data, delivered through Snowflake, MCP, or flat file.