从“效率工具”到“价值引擎” 大模型“再造”银行

Core Insights - The banking industry is transitioning from pilot AI applications to comprehensive integration, driven by the launch of domestic large models like DeepSeek-R1 and V3.1, with over 330 applications already implemented by Ping An Bank by June 2025 [1][2] - Banks view AI as a key strategy for enhancing operational efficiency, optimizing customer experience, and exploring new growth avenues, shifting from cost-cutting tools to revenue-generating engines [1][2] Industry Application Acceleration - Ping An Bank has developed a large model capability system, enhancing application capabilities through various advanced techniques, with over 330 applications in place by June 2025 [1] - The bank has deepened the use of intelligent algorithms in risk control, establishing a risk expert knowledge base and improving efficiency across risk management processes [2] - Shanghai Bank emphasizes the need to integrate its services with large models to enhance digital transformation efforts [6] Business Model Reconstruction - The banking sector is experiencing a significant evolution in large model applications, moving from efficiency tools to value creation engines, although challenges remain in controllability, interpretability, and cost-benefit ratios [6][7] - The maturity of large model applications in banking shows a tiered structure, with some applications nearing maturity while others are still in exploratory phases [7] - Jiangsu Bank has successfully implemented a large model for business material entry and review, significantly improving operational efficiency [8]