On August 28th, China Minsheng Banking Corp. held its 2025 interim report performance conference. The bank's Chief Information Officer Zhang Bin introduced the institution's work in digital transformation and AI initiatives during the first half of the year.

Zhang stated that AI represents the most important driving force behind paradigm transformation in technology and banking operations. The bank places great emphasis on this area, continuously enhancing systematic capabilities in the first half while expanding business scenario applications in both depth and breadth. During H1, the bank achieved deployment of its AI agent platform and added 78 detailed application scenarios, compared to 80 scenarios for the entire previous year. This demonstrates that Minsheng Bank's AI applications, including large language models, are showing accelerated momentum.

This acceleration stems from two aspects. First, the bank consistently focuses on building enterprise-level capabilities while maintaining effective organizational promotion. Additionally, by the end of June 2025, the daily average call volume for large model applications increased threefold compared to year-end, reflecting the expansion of applications in both depth and breadth.

Second, during the first half of this year, scenarios involving large models primarily included credit investigation, retail marketing, remote customer service, business operations, and knowledge Q&A. In credit investigation, large models assist in generating various risk analysis reports, enhancing customer risk analysis and forecasting capabilities and efficiency. Currently, 80% of new credit scenario functions have been implemented. Large models are used to assist retail account managers with marketing, including customer profiling and wealth management recommendation functions, which have been promoted bank-wide. Through large models for intent recognition and knowledge base construction, the problem resolution rate of the mobile banking app's chatbot improved by 16 percentage points.

Zhang concluded by stating that applying large models to the entire IT and R&D process and embracing Software Engineering 3.0 represents an important strategic initiative in Minsheng Bank's digital transformation. Previously, the bank primarily used large models for code generation, unit testing, and interface development. In the first half of this year, they expanded to the full technology lifecycle, including requirements development, quality inspection, testing, and operations across all domains. The bank is also exploring the construction of R&D AI agents and end-to-end AI enablement.