AI driven strengthening digital payments with intelligent analysis

With the rapid increase in UPI-based digital transactions, fraudulent activities have also risen significantly. Conventional fraud detection frameworks often resort to unchangeable, regulations-centric approaches that are inadequate for recognizing intricate or developing fraud schemes. Hence, this study suggests employing machine learning methods like Random Forest and XGBoost to develop an intelligent fraud detection system that can uncover abnormal activities. Along with the amount, the device id, and the number of failed attempts; the system takes a transaction, detects a fraud instantly, and even gives a visual confirmation. The model predication accuracy user interface responsiveness and fraud identification rate 90%. Bulk fraud detection with SVM, LSTM 91% and compare machine learning XGBoost fraud detection 95% accuracy. The model is deployed via a web interface supporting both individual and bulk transaction analysis. This approach enhances payment security and builds user trust in digital platforms.