Claims assessment and underwriting
Each assessment compares Sentinel-1 radar and Sentinel-2 optical scenes before and after the event and returns a result with a confidence indicator, or an explicit "insufficient data" with the reason.
SAR and NDVI windows either side of the event date, with the affected area and its error margin.
In practice: A storm flattens part of a field — the assessment weighs the radar change and the NDVI drop and returns the damage with an interval.
Damage percentage and affected hectares returned as GeoJSON zones, not a single flat number.
In practice: A large field with a partial loss — the zones show which part was hit, so you can settle on the damaged hectares.
PDF with a sha256 hash and issue time; the hash is stored with the assessment, so any copy can be checked against it.
In practice: A farmer disputes the assessment a year later — the regenerated certificate has the same hash, so the evidence stands.
A risk class from 20 years of drought history, yield variability and flood exposure, and parametric triggers evaluated per year with the scenes kept as evidence.
In practice: Quoting a new policy — you see the parcel’s risk class (A–E) from its own hazard history, not from claims alone.
Available as an API
Order assessments with a claim reference and get the result by webhook; the free sandbox has five modelled damage cases.
Access to the RISK workbench is issued by Agrovisio. Name a storm or drought that already hit your book: we assess it retrospectively and you compare our output against your settled claims.
Please include the event, the date and the number of claims in scope.