The project “AI-powered remote monitoring model for detecting operational risks of agency financial services”, developed by Sichuan Postal Branch over the past year, has been officially approved by China Post Group.

Since the launch of the project, Sichuan Postal Branch has conducted pilot validation in three phases. Through over 200 algorithm iterations and more than 5,000 times of testing optimization, it has created an AI-powered daily review algorithm model to identify 15 irregular scenarios, including proxy operation of customers’ mobile phones, proxy operation of ITM (Interactive Teller Machine), and selling products outside surveillance areas. In addition, Sichuan Postal Branch has independently developed an AI-powered early warning management system and constructed a database of 75 algorithm scenarios in 15 categories. It has also standardized the hardware and software configuration for AI-powered detection and formulated corresponding AI monitoring regulations. Eventually, from the ground up, it has built a complete AI-powered daily review closed-loop management system from irregularity detection, warning verification, risk handling, warning notifications, risk analysis, to data feedback.