ClaimFlow cut its claims cycle from 12 days to 36 hours.

ClaimFlow is a Midwestern Property & Casualty carrier writing $1B in annual premiums across 250,000 active policies. They serve auto, home, small commercial, and specialty lines through a network of 600 independent agents across nine states.

Two years of 30% growth had stretched the team thin. The operations behind the policies — claims intake, document review, fraud triage, settlement workflow — still moved on paper and email, and the gap between what the business was selling and what the back office could service was getting wider.

ClaimFlow’s claims operation was 30% bigger than two years ago, but the back office still ran on paper, email, and an aging claims system. Three failures showed up everywhere we looked.

We were paying senior adjusters to key in photos and forms — and making customers wait days for a decision a checklist could make.

We rebuilt the claims pipeline end-to-end — replacing handoffs with one automation layer that thinks. Documents are read. Fraud signals are scored at intake. Simple claims resolve themselves. Complex ones reach the right adjuster with everything already pre-filled.

A fraud-triage orchestrator sits at the centre: 30+ business rules layered with an XGBoost model score every claim in seconds, with the reasons surfaced for adjusters rather than hidden in a black box.