Insurer Settles Storm Food-Spoilage Claims in Hours, Not Days
Quick Summary
Allianz Australia built a team of narrow AI agents that handle small food-spoilage claims after storms: one checks coverage, one checks that the weather event really happened, one screens for fraud, one calculates the payout, one audits the rest. Processing time fell 80%, from several days to a day or even hours.
The Use Case
After a storm, thousands of customers file claims for spoiled food under $500. Each one is simple, but the volume buried adjusters and pulled them off complex claims. Allianz Technology built "Project Nemo" in under 100 days. A planner agent starts the workflow, a coverage agent confirms the policy includes food spoilage, a weather agent confirms a matching event at the customer's address, a fraud agent checks for red flags, a payout agent calculates the amount, and an audit agent reviews every decision and writes a summary. The automated steps finish in under five minutes. A person still makes the final payout call.
Real Example
- A customer submits a $340 food-spoilage claim two days after a cyclone
- The coverage agent confirms the policy covers it
- The weather agent matches a power outage and storm at the address on those dates
- The fraud agent finds no red flags; the payout agent calculates $340
- The audit agent writes a one-paragraph summary of every check
- A claims handler reads the summary and approves payment the same morning
Why It Matters
- Processing and settlement time down 80%
- AI steps complete in under 5 minutes before human review
- Built and launched in under 100 days