Field notes
Claims operations, in the open.
How we think about AI agents for insurance — fraud detection, human-in-the-loop approvals, and the messy reality of running claims at scale.
Claims are edges, not rows
Why Grivara models a claim as a typed edge instead of a row — and how a three-layer ontology (conceptual, contract, provenance) makes every decision auditable.
#engineering#graph#product#reference· 10 min readWelcome to the Grivara blog
Why we're building an AI claims command center — and what we'll write about here.
#company#product· 4 min readThe fraud landscape got synthetic
Generative AI made fabricated evidence cheap, and rule-based defenses built for the pre-AI era can't keep up. Here's what broke and how Grivara closes the gap.
#fraud#product#engineering· 9 min readInside graph collusion
What a 3-hop neighborhood actually tells you about a fraud ring — and how we turn it into a single number an adjuster can trust.
#fraud#graph#engineering· 8 min read