Detects and analyzes transactional fraud using a multi-model approach with large language models and external tools.
This project provides an advanced assistant for fraud analysis, leveraging large language models and a suite of external tools. It's designed to scrutinize transactions for anomalies, aggregate crucial threat intelligence, and deliver clear, explainable risk scores to investigators, streamlining the fraud detection and risk management process.
Key Features
01Inspects transactions for anomalies
02Aggregates threat intelligence
03Generates explainable risk scores
04Multi-model fraud detection using LLMs and external tools
050 GitHub stars
06Handles PII via mode-switching and redaction
Use Cases
01Consolidating threat intelligence for comprehensive risk assessment
02Providing detailed risk score explanations to fraud investigators
