Azure reference architectureClaims intake and processing
A Azure reference architecture for claims intake and processing in Insurance. Triggered by first notice of loss (fnol) received, Padiso orchestrates 3 specialist agents on Azure AI Foundry, reaching policy admin system and your tools through the Model Context Protocol, with budgets, approvals and an audit trail throughout.
AzureRuntime: Azure AI FoundryTrigger
First notice of loss (FNOL) received
Insurance
Orchestration
Padiso Orchestrator
goals · tasks · heartbeat
Governance
budgets · approvals · audit
Runtime & agents
Agent runtime
Azure AI Foundry
Claims agent
agent
Underwriting agent
agent
Servicing agent
agent
Integrations · MCP
MCP gateway
Your systems
Policy admin system
Claims system
Fraud / scoring
How the data flows
Trigger: First notice of loss (FNOL) received
- 1
Claims agent intakes the FNOL and validates policy coverage
- 2
It gathers documentation and runs anti-fraud checks
- 3
Straight-through low-risk claims; the rest route to an adjuster with a score
- 4
Payouts require human approval before disbursement
Agent team
- Claims agent
- Underwriting agent
- Servicing agent
Integrations
Reached through the Model Context Protocol, scoped and governed.
Runtime & data
Azure AI Foundry
Policy admin system
Claims system
Fraud / scoring
Security & operational considerations
- Anti-fraud rules run before any payout
- Approvals mandatory on disbursement
- Full audit trail for regulators
Azure deployment
Deployed in your Azure tenant on AI Foundry, with Entra ID scoping, Private Endpoints into your VNet, and Azure Monitor for audit and observability.
Run this blueprint anywhere
More Insurance blueprints
Run your business on agents.
Start free, or talk to us about putting an agent workforce to work across your operations.