Building an Agentic Fraud Investigation Agent with TigerGraph, MCP, and GraphRAG

FraudGraph AI is an agentic fraud investigation platform developed using TigerGraph, MCP, and GraphRAG. Moving beyond binary classification and raw anomaly scores, the system focuses on the relationships between customers, cards, devices, and historical cases. It combines graph-based evidence, agentic tool use, and deterministic fraud policies with persistent case memory. This approach allows analysts to gather supporting evidence and assess uncertainty, ensuring that investigation actions strictly follow established banking policies rather than relying on isolated transactions.