Allianz Project Nemo: 7 AI Agents for Insurance Claims and Fraud Detection

Allianz deployed Project Nemo, a multi-agent AI system with 7 specialized agents for insurance claims automation and fraud detection. The system achieved 80% reduction in claims processing time. For food spoilage claims under AUD 500, processing time dropped from several days to hours. The system processes 15,000 claims daily with 94% accuracy. Human adjusters now focus exclusively on the 6% of claims flagged for complexity. Allianz Switzerland also stopped over EUR 12 million in fraudulent payouts using AI-driven fraud detection. The multi-agent architecture uses a triage agent for initial claims routing, a precedents search agent, and a decision agent that recommends payouts based on coverage terms and fraud risk signals.