- Sub-tier depth and supplier burden:
- Map a real product through every tier and record where supplier data stops and what is inferred. The pages claim reduced supplier fatigue, so measure response effort with actual suppliers.
- Geolocation validation test:
- Upload a file with swapped coordinates, open polygons, overlaps and low precision, and record what is blocked. The reviewed pages claim validation without stating rules, so require a written specification.
- Risk model transparency:
- Ask which factors, weights and satellite data produce a score and whether thresholds can be customised, as the launch post suggests. Keep an example scoring record and the change history of any adjusted model.