Public banks' 2025 "report cards" are gradually being released. What trends in AI applications in the financial industry are being revealed?

China National Radio Beijing, March 28 (Reporter Mu Di) “Comprehensively promote in-depth integration of artificial intelligence and featured businesses, and accelerate the rollout of application scenarios such as intelligent wealth management, carbon-neutrality management, and multi-asset quantitative strategies,” Industrial Bank mentioned in its 2025 annual report released recently.

The China National Radio Finance reporter noted that the “report cards” of listed banks for 2025 have been gradually coming to light. In the midst of rapid advances in artificial intelligence, many banks have discussed the development of financial technology capabilities in their annual reports, earnings conference calls, and other settings, including topics related to AI technology development and applications.

Analysis indicates that listed banks pushing ahead with “artificial intelligence + finance” marks a new stage in the industry’s digital transformation—from “informatization” to “digital intelligence.” Looking ahead, in building financial technology capabilities, commercial banks should do a good job in areas such as compliance and risk control, data governance and standardization, and talent team building.

Ping An Bank: “Artificial intelligence + finance” is one of the core strategies

Recently, at Ping An Bank’s earnings conference call, Deputy President and Chief Financial Officer Yiou Zhiyi stated that applications of AI in the financial industry have attracted great attention. In the new round of the technological revolution era, “artificial intelligence + finance” is also one of Ping An Bank’s core strategies.

“This work is forever unchanging—two aspects,” Yiou Zhiyi said. On one hand, it is the development of technology capabilities; based on today’s artificial intelligence, Ping An Bank needs to strengthen AI’s technology and data—these are the two major capability foundations. In the application portion, Ping An Bank mainly focuses on digital employees, precise marketing, and precise risk control, taking these three as the main levers for applications. Ultimately, it is to enable intelligent operations, intelligent management/management for operations, and intelligent risk control, including management, to play an even greater role.

Yiou Zhiyi said that these areas have never stopped in the past; especially over the past one or two years, the bank has been gradually investing in these areas, and more achievements have been demonstrated. For example, on cost reduction: in retail business, there is currently an AIGC marketing platform. On this platform, many creative content pieces are automatically generated by digital humans, reducing a lot of manual work. In rough terms, last year this business segment cut expenses for Ping An Bank by about 60 million yuan. He said that overall, Ping An Bank is definitely continuously iterating and upgrading the capability foundation by strengthening its technology capability building. The ultimate purpose of capability building is to be applied to actual marketing, operations, and risk management processes, and finally achieve cost reduction, quality improvement, and efficiency gains.

China CITIC Bank: Technology capabilities are accelerating into core business processes

According to an incomplete review by the reporter, building financial technology capabilities—including artificial intelligence—has become one of the major directions commercial banks are pushing.

“In recent years, our bank has steadfastly advanced the implementation of the ‘leading digital bank’ strategy,” at an earnings conference call, Gu Lingyun, Deputy President of China CITIC Bank, introduced that overall, the bank’s technology capabilities are accelerating into core business processes, releasing real productive forces in four areas: reducing costs, improving efficiency, controlling risks, and providing better user experience.

Gu Lingyun gave examples. In terms of financial market business, relying on the bank’s independently developed centralized quantitative trading platform, the overall automation rate of trading has exceeded 80%. The bank also was among the first to introduce large-model-enabled generation of quantitative strategies; the strategy R&D efficiency has improved by 3 times, accurately capturing cross-market price spreads in international gold, and enabling fully automated order placement through millisecond-level quantitative algorithms—effectively avoiding delays caused by manual operations. Daily average trading volume has reached more than 10 times that of 2024. The AI quotation robot has covered four major markets: money, derivatives, cash bonds, and foreign exchange, improving efficiency by more than 5 times and driving a 25% increase in trading volume. Meanwhile, by using the bank’s independently developed valuation algorithms, it can independently value major derivatives such as FX options and structured deposits, significantly reducing reliance on overseas models, saving over one million yuan in costs every year.

Gu Lingyun introduced that in terms of AI applications, in 2025 the bank’s investment in intelligent computing hardware increased 5 times. It has already built a unified AI middle platform and a GPU computing power cluster. Currently, the monthly call volume of small models is close to 500 million times, and the daily peak call volume of large models exceeds 3 million times.

Looking ahead to 2026 and the “15th Five-Year” period (the 2026–2030 timeframe), Gu Lingyun said that China CITIC Bank will continue to anchor the “strengthen technology by force” goal, striving to build an industry-leading “digital intelligence bank,” focusing on four things: first, make organizational operations run more smoothly. Second, make data more effectively utilized. Third, integrate artificial intelligence into every corner. Fourth, make the security barrier even more solid.

AI is seen as a key driving force for structural change

When discussing listed banks’ push into “artificial intelligence + finance,” Tian Lihui, a professor of finance at Nankai University, told the China National Radio Finance reporter that this signifies that the banking industry’s digital transformation has entered a new stage—from “informatization” to “digital intelligence.” This is inevitably a response to the narrowing net interest margin and operational pressure. By using AI to reduce costs and increase efficiency and to enhance the precision of risk control, it is also a strategic positioning to reshape core competitiveness.

Tian Lihui said that from related cases, it can be seen that banks no longer view AI merely as an auxiliary tool “adding icing on the cake,” but rather as a production engine fully embedded into core business processes. This trend will profoundly change banks’ cost structures, service models, and competitive landscape. In the future, differentiation will depend on who can truly transform AI capabilities into sustainable profitable growth.

“Currently, listed banks are treating AI as the key driving force for structural change and continually increasing investment in AI and its applications,” said Lou Feipeng, a researcher at China Post Savings Bank. Based on the actual situation, banks have massive amounts of data, which gives them an advantage in AI applications. AI can significantly reduce unit operating costs and open up new revenue sources through intelligent risk control, automated customer service, precise marketing, and more, thereby driving the banking industry to transition toward an intelligent, light-capital model.

Looking forward, Tian Lihui advised that commercial banks accelerating AI capability building should focus on four dimensions. First, strategically, “do something and refrain from doing something”—AI investment is huge, and it must be concentrated on core scenarios that can generate the most value, such as risk control, marketing, and operations, rather than blindly pursuing a “big and complete” technology stack. Second, data governance is the foundation; the precision of AI models highly depends on data quality, so banks must break down internal data silos, and establish unified data standards and a governance system. Third, technology and business must be deeply integrated, avoiding the “two separate layers” phenomenon; technology personnel should understand business, business personnel should understand AI, and an agile iteration mechanism for coordinated collaboration should be formed. Fourth, risks and compliance cannot be neglected; issues such as the “black-box” nature of AI models, algorithm bias, and data security need to be controlled upfront, and an explainable and auditable AI governance framework should be established. It should anchor strategic synergy, scenario value, security bottom lines, and a talent ecosystem, take serving the real economy as the orientation, and ensure that AI truly transforms from a “technology highlight” into “business performance.”

Lou Feipeng also told the reporter that in building financial technology capabilities, commercial banks should, first, strengthen compliance and risk control, and make effective use of the advantages of technology under compliance. Second, strengthen data governance and standardization, break down information silos, and improve the efficiency of anti-fraud data flow. Third, strengthen talent team building, and do a good job in organizational and institutional transformation.