Designed a first-notice-of-loss voice workflow that captures structured claim details, documents evidence, and speeds up adjuster review.
Challenge
An insurance operation needed faster first-notice-of-loss intake after accidents, storms, and property incidents. Human agents collected inconsistent information, photo evidence arrived through scattered channels, and adjusters often had to call customers back for missing details.
Workflow approach
We built a claims intake voice agent that guides callers through the right question path based on claim type. It captures parties involved, incident location, timeline, damage details, policy identifiers, photos, and preferred contact channels. Urgent or sensitive cases are escalated immediately, while standard claims are packaged for adjuster review.
Observed outcome
Intake time fell 58%, complete claim packets rose to 83%, and adjusters spent less time chasing missing information. Customers could report claims after hours without waiting for a live agent.
Outcome metrics
- Faster intake
- 58%
- Complete claim packets
- 83%
- FNOL availability
- 24/7
- Less adjuster rework
- 31%
Systems involved
- VAPI
- GPT-4o
- Twilio
- S3
- Claims API
- n8n