FNAL AI

Claims intake is a flood of unstructured detail: descriptions, photos, follow-up questions, all arriving while adjusters juggle dozens of active cases. Important signals get buried in the thread.

Designed AI-assisted surfaces that summarize, extract, and suggest while keeping the adjuster visibly in control. The hard design problem was trust: making machine output inspectable and correctable rather than authoritative. Shipped capabilities include claim summarization, writing assistance, fraud and risk signals, and image analysis, each surfaced as suggestions the adjuster accepts, edits, or dismisses.

AI assistance integrated into core carrier workflows with human-in-the-loop controls, reducing time spent synthesizing unstructured claim details.