ClaimVista
Reduce claims cycle time and improve adjuster productivity with AI-powered triage and data extraction.
ClaimVista is claims automation that reads first notices of loss, photos and estimates as they arrive, extracts the policy and loss data, scores each claim for complexity and fraud risk, and hands adjusters a pre-filled worksheet or settles the simple ones outright.
Who is this for?
Claims operations leaders in auto, property, workers' compensation and health, and the adjusters who spend the first ten minutes of every claim typing. It is for carriers on Guidewire ClaimCenter, Duck Creek Claims or a legacy platform, with a CAT season somewhere in the calendar.
What problem does it solve?
The intake step is the slow one. A FNOL arrives with photos, a police report and a repair estimate, and someone keys it into the system field by field before anyone can judge it. Simple claims wait behind complex ones because both landed in the same queue, and the fraud indicator, if it was there, is noticed after the reserve is set.
During a storm the volume multiplies and the headcount does not. Contractor estimates vary, coverage has to be verified against policy terms by hand, and the cycle time that was already 18 days stretches.
How does it work?
Documents arrive by any channel. OCR runs at 95 percent or better on structured forms such as FNOLs, police reports and repair estimates, and at 90 percent or better on field photos, with Mitchell, CCC and Audatex estimate formats read natively; low-confidence fields are highlighted for the adjuster rather than guessed. A triage model then scores each claim by complexity, urgency and fraud risk and routes it: simple claims to junior adjusters or straight-through processing, complex ones to senior staff, flagged ones to an SIU queue.
Adjusters get a single screen with every document, every extraction and its confidence score, and a worksheet already filled in, which saves eight minutes or more of data entry per claim. Fifty-plus red-flag rules watch for late reporting, excessive treatment, prior claim history, provider patterns and duplicate losses across channels.
The automation level is yours. Straight-through processing can settle 30 to 40 percent of simple, low-risk claims with validated data, no fraud flags and within policy limits; everything else stays human in the loop with 80 percent of the data pre-extracted. Fraud scoring is tunable and never denies on its own; precision and recall sit on the KPI dashboard.
Capabilities
- Auto-triage claims by complexity, fraud signals, and urgency
- OCR extraction of policy/loss data from photos, PDFs, and forms
- Pre-filled adjuster worksheets with validated data
- Fraud signal detection with 50+ red flag rules and SIU queue routing
- Integration with core claims systems (Guidewire, Duck Creek, etc.)
Outcomes
- Cycle time ↓ 30-50%
- Data entry ↓ 8 min saved/claim
- Extraction accuracy 95%+
What does it connect to?
- Claims platforms: Guidewire ClaimCenter (two-way claim sync, task assignment) and Duck Creek Claims (policy and claim data exchange, workflow triggers)
- Estimate formats: Mitchell, CCC and Audatex
- REST API and webhooks for proprietary systems
- SIU review queue and the KPI dashboard
How is it deployed, and how long does it take?
Cloud for most carriers, on-prem where the data residency rule says so; HIPAA-compliant processing with a full audit trail for health and workers' compensation lines. Change management for adjusters is part of the pilot, because a pre-filled worksheet only saves time if people trust it.
Reserve recommendations and treatment summaries are suggestions with the evidence attached, not decisions; the adjuster still owns the file.
Deployment in 60 to 90 days. Payback usually within one quarter through lower loss adjustment expense, with cycle time down 30 to 50 percent.
Industry context
Explore the related workflows and other solutions for your industry.