Claims agent with limits
An agent may settle simple claims below a set amount; anything above goes to a claims handler.
Consumer finance
Insurance underwriting and claims
Pricing and claims decisions made with AI need oversight and a record of what was decided and why.
Insurers use models for pricing, underwriting and claims triage, and agents increasingly gather documents and propose settlements. In the EU, the AI Act lists AI systems used for risk assessment and pricing in life and health insurance for natural persons as high-risk. AgenticOpenFinance™ adds mandates, human review and an evidence trail around every agent action that touches a policyholder.
What agents do
What you build on AgenticOpenFinance™
Document gathering under consent
The agent reads only the records the policyholder consented to, for the stated purpose and period.
Human review queue
Escalated decisions reach a handler with the agent's proposal, the inputs and the policy that fired.
Payout under mandate
Approved settlements are paid to verified accounts only, with the payee check recorded.
Decision records
Each decision keeps the model version, inputs, verdict and reviewer, as automatic logging would require.
Value chain
The actors in this chain
AgenticOpenFinance™ connects them with one identity check, one mandate model, one policy and one evidence trail.
PolicyholderInsurerClaims agentClaims handlerLoss adjusterPayment provider
Typical flow
A governed agent flow, step by step
- Claim submitted
- Agent gathers documents under consent
- Agent proposes outcome
- Policy check and threshold
- Handler reviews if escalated
- Payout to verified account
- Evidence sealed
Live demo
See it in the terminal
A multi-agent invoice workflow with a policy verdict and the finance lead's approval (demo). Demo data
DemoAll companies, people, agents and amounts are fictional. Protocol names refer to public specifications; naming one is not a partnership or an endorsement.
Glossary