Insurance
Multi-Party Vehicle Repair Claim Orchestration
A leading auto insurance provider faced a massive bottleneck in the post-approval phase of vehicle damage claims. Coordinating between the policyholder, the assigned repair garage, and the field surveyor involved days of manual follow-ups, phone tag, and scattered emails. This led to extended repair times, breached SLAs, and a flood of inbound "Where is my car?" calls to the contact center.
The Specific Pain Point
Claims managers were acting as highly-paid dispatchers, manually chasing garages for updates and attempting to align the schedules of independent field surveyors with the garage's unpredictable repair completion times.
The Orchestration Flow
Subverse deployed a Claim Lifecycle Orchestrator. Once a claim is approved, the orchestrator triggers three distinct agents:
Agent A (Garage Coordinator) proactively pings the repair shop via WhatsApp at scheduled intervals to request status updates and proof-of-repair.
Agent B (Surveyor Scheduler) monitors Agent A's progress. Once the garage confirms completion is nearing, Agent B cross-references surveyor calendars and automatically schedules the final post-repair inspection.
Agent C (Customer Liaison) translates these backend movements into proactive, plain-language updates sent via SMS to the policyholder. If a garage is delayed, the Orchestrator instantly reschedules the surveyor and notifies the customer without human intervention.
System Integrations
Guidewire (Core System), WhatsApp Business API, internal surveyor Outlook/Exchange APIs, Garage Management Portals.
Measurable Outcome
Claim lifecycle reduced from 14 days to 6 days. Inbound ETA calls to the contact center dropped by over 40%. Claims managers track the entire network's performance via Subverse's conversational reporting dashboard.
Multimodal & Memory Elements
The flow utilizes multimodal inputs, allowing the garage to upload photos of the repaired vehicle via WhatsApp, which the AI pre-screens for clarity before notifying the surveyor. Entity-level memory tracks historical garage performance; if a specific vendor consistently runs two days late, Subverse automatically adjusts its proactive check-in schedule for that specific garage to prevent future delays.
