LLM-Powered UPI Transaction Monitoring and Fraud Detection

In this Business World, the Unified Payments Interface (UPI) has transformed digital payments by facilitating instantaneous Business-to-Person (B2P) and Person-to-Person (P2P) transactions. But the quick expansion of UPI transactions has also drawn fraudsters. Current fraud detection systems are built on Machine Learning algorithms that have been trained on historical data and Rule-Based methodologies. Even if they work, these techniques can't keep up with changing fraud trends and can overlook new attack opportunities. The inflexibility and absence of Real-Time Natural Language understanding in current systems makes it difficult to assess user behaviour and spot false narratives. This study uses Large Language Models (LLM's) to provide a unique method of fraud detection and UPI transaction monitoring. After being trained on enormous text and code datasets, LLM's are able to discern suspicious patterns and narratives that point to fraud by analyzing transaction data, user messages, and social media interactions.