Earnings calls dominate the headlines as the banking industry's AI competition enters deep water

When trading stocks, rely on Golden Kirin analyst research reports—authoritative, professional, timely, comprehensive—helping you uncover high-potential theme opportunities!

Source: Beijing Business Daily

Back when loan services required lining up for hours and filling out complicated forms, today you just tap a mobile banking app, and within minutes you can get accurate, targeted notifications of credit limits and interest rates. At bank branches, tellers and AI assistants hold real-time conversations to crack tough business problems—“silicon-based colleagues” have become indispensable work partners for frontline staff. From state-owned big banks to local banks, from top-level strategic planning to frontline scenario implementation, AI is penetrating from every angle, driving banks to leap from “rule by people” to “rule by technology.” At the recently held, tightly packed earnings press conferences of major banks, executives repeatedly mentioned AI strategy—an industry transformation led by AI has already been fully rolled out. This transformation not only reshapes banks’ business models and collaboration methods, but also pushes financial services back to their essence, extending the service radius indefinitely and continuously refining service granularity. However, opportunities and challenges coexist: as the AI race enters deeper waters, problems such as data silos, privacy protection, and regulatory adaptation still urgently need to be solved.

From “tool workers” to “silicon-based colleagues”

Against the backdrop of continuously rising complexity in bank operations, AI assistants have become the key to breaking through “knowledge silos” and boosting frontline service capability. At the 2025 annual earnings conference, Lei Ming, deputy chairman of the China Construction Bank, disclosed a set of data: as of the end of 2025, during the process of responding to branch inquiries at China Construction Bank, the coverage of AI assistants had already reached 99.42%, and the average daily visit volume had exceeded 100,000 person-times.

This means that when employees encounter business problems and seek help from the head office or management departments, in most cases, artificial intelligence will be the first to provide solutions. This “strongest brain” super mentor stands by at all times, extremely patient, operating year-round without rest—changing the internal collaboration model of the bank.

Within the Industrial and Commercial Bank of China’s “ICBC Intelligent Surge” large-model technology framework, changes are also especially notable. The bank has already scaled up and deployed more than 500 AI applications across over 30 business areas, and its AI digital employees annually handle the equivalent of 55,000 person-years of work. These “employees” that require no payroll can work nonstop 24/7, sharing the massive pressure of business operations. China Merchants Bank uses large-model technology to improve the quality and efficiency of financial expense reimbursement. By the end of last year, it processed 1.4085 million paperless reimbursement slips, up 23.76% year over year. Industrial Bank’s AI programming assistants cover 90% of R&D personnel; Morning Evening Meeting’s intelligent agent assistant covers more than 1,500 department and institutional branch outlets.

AI has become an indispensable “work partner” for employees across every line of banking. China Merchants Bank, on the retail side, has built a series of retail “little assistants,” continuously empowering relationship managers and the middle/back office teams in scenarios such as customer operations, operational analysis, and wealth management and research/asset allocation. On the wholesale side, it has built “CRM little assistants” to help corporate relationship managers improve service quality and efficiency. On the risk side, it has built “risk little assistants,” embedding them into operating workflows to enable intelligent risk management. On the operations side, it has built “operations little assistants,” enabling scenario applications such as digital assistant functions, Q&A for operational knowledge, intelligent business review/verification, intelligent service practice, and intelligent analysis of risk events. As of the end of 2025, user coverage for corporate relationship managers, credit personnel, and operations personnel using the corresponding “little assistants” reached 80.13%, 80.32%, and 100%, respectively.

On the frontlines facing customers directly, AI is redefining the boundaries of “service.” Traditional bank services are constrained by labor costs and often cannot truly deliver personalized experiences. But now, for example, Bank of Communications has added AI product interpretation and AI-assisted generation of research and investment views into its wealth management system, meeting broad customers’ needs for personalized asset allocation. Ping An Bank upgraded its “AI+T+Offline” service model, strengthening the use of digital tools such as AI assistants and intelligent outbound calling to improve the efficiency of remote banking services. CITIC Bank leverages small-model + large-model capabilities to empower the back-office handling and consolidation of corporate account opening, changes, and other businesses, fully building a new operating model and boosting the effectiveness of business centralization by more than 2 times.

At the 2025 annual earnings briefing, a statement from Lu Jiajin, chairman of Industrial Bank, further pinpointed the future trend. In his view, “In the AI era, silicon-based life will replace a large amount of work done by carbon-based life, and you can feed financial-related knowledge to some AI agents—covering funds, retail, and interbank—so one person can play multiple roles separately. Going forward, relationship managers will no longer be divided by types such as company clients, retail, and interbank.”

AI goes “into the bank,” penetrating across the entire domain

The core logic of this AI “penetration battle” is that banks are crossing the gap from traditional “rule by people” to efficient “rule by technology.”

From the “aircraft-carrier” turnaround of state-owned big banks, to the agile breakthroughs of joint-stock banks, and then to the precise positioning of local banks, AI is no longer just a nice-to-have embellishment—it has become the nervous system that penetrates into the business’s capillaries.

At the level of top-level design, major banks are placing their bets. Based on the latest data disclosed in their annual reports, Industrial and Commercial Bank of China implemented its “Leading the AI+ Action” plan at the group level in 2025, empowering four core scenarios: investment trading, marketing and customer acquisition, risk control, and operational efficiency improvement. Postal Savings Bank of China opened 10 categories and 24 general AI capabilities to each of its branches, forming its “AI2ALL” digital ecosystem with “full-domain outreach to the outside” plus “companywide efficiency enhancement from within.”

China Merchants Bank proposed the “AI First” concept. On this bank’s strategic chessboard, AI is placed in a position of “priority, leading, and being first.” Changes in top-level design also determine the flow of resources. Whether it is the “little assistant” for the retail line or the “little assistant” for the wholesale line, AI no longer waits for business needs—it proactively embeds into and reshapes business processes.

Local banks are also not lagging behind. Multiple banks that have already disclosed their annual reports have also devoted significant focus to AI strategies. Chongqing Bank has built the “Chongyin Xia oAI” brand application, becoming one of the first batch of city commercial banks nationwide to achieve large-model “private deployment + financial scenario adaptation.” Qingdao Bank has drafted the “Qingdao Bank Digital Transformation New Three-Year Strategic Plan,” which mentions building two major intelligent engines for AI capabilities and data value. Rural Commercial Bank of Yifeng has also made it clear that in 2025 it will build a whole-bank AI platform based on open-source frameworks, forming an intelligent agent application ecosystem covering major business lines, and that AI capability building has entered a stage of scaled-up applications.

AI has also become a high-frequency term at earnings press conferences. Looking ahead, focusing on the key tasks of building a “digital-intelligent Industrial and Commercial Bank of China,” ICBC deputy chairman Zhao Guid e pointed out that it will continue to implement the “Leading AI+” action, focusing on four areas: intelligence, smartness, intelligent computing, and intelligent enjoyment. It will innovate to build financial intelligent agents, shifting technology positioning from behind-the-scenes support to front-stage driving. It will accelerate the construction of a “one customer, one advisor” service model, so that AI becomes the most direct bridge connecting the bank and customers.

Bank of Communications deputy chairman Qian Bin stated clearly that it will drive AI to transform from single-point applications to full integration. His proposed moves—strengthening the bank’s own technological capability building, deepening services between business operations and employees, upgrading the service market and customer experience, and improving the level of intelligent risk prevention and control—make it clear that AI has been deeply embedded into the bank’s top-level design, becoming new-quality productive forces that drive cost reduction, quality improvement, and efficiency gains.

Wu Zewei, a special research fellow at Sushang Bank, pointed out that AI’s autonomous decision-making, real-time responsiveness, and intelligent learning capabilities will comprehensively reshape banks’ business models. This includes upgrading customer experience: AI can redefine how banks connect with customers through multimodal interactions and personalized services, enabling full-cycle customer companionship, personalized wealth management, and real-time anti-fraud monitoring. It also includes upgrading risk management: AI can shift risk control from “after-the-fact response” to “real-time interception + predictive early warning,” enabling innovations in credit assessment, identification of complex fraud, and compliance automation, thereby building an all-process protection network. It also includes upgrading operational efficiency: AI drives banks’ business workflows to evolve toward “zero-touch” and “adaptive,” releasing organizational productivity and achieving process automation, scientific decision-making, and the evolution of organizational knowledge.

These challenges still need to be solved

From architecture evolution, to full integration, and then to smart decision-making, the AI competition in the banking industry has entered deeper waters.

How to make technology applications safer and more controllable has become the foremost consideration in the banking industry’s digital-intelligent transformation. In the earnings meeting, ICBC president Liu Jun explicitly said that the prerequisites for technology applications are key. He stated, “The technology used by Industrial and Commercial Bank of China is relatively new, but this technology must be validated by the market and by our strong internal validation capabilities; otherwise, we would not dare to place this technology on top of the system in a rush, because protecting customer privacy and information security is the bank’s most important responsibility.” Liu Jun emphasized, “Therefore, Industrial and Commercial Bank of China will integrate advanced technology into operational workflows, and it must have system validation as a prerequisite.”

CITIC Bank deputy chairman Gu Lingyun also emphasized, “Make the security barrier even more solid, lay intelligent computing capacity ahead of time in an appropriate manner, introduce new types of security technologies, and ensure that AI applications are safe, trustworthy, and controllable.”

Zhao Guid e also discussed improving governance effectiveness and building an end-to-end security and prevention/control system for AI applications, effectively covering areas such as the security of technology infrastructure, data security, model security, and application security.

In Dong Shimiao’s view, chief researcher at Zhaolian, AI applications not only drive positive change at the business, organizational, and cognitive levels, but also bring new problems in technology, regulation, and talent. On the technology side, the “data silos” formed within a fragmented data ecosystem can cause model bias, and protecting data privacy and security during training is also a pressing issue. On the algorithm side, the opaque model decision-making process and the risk of “hallucinations” in generative AI increase the difficulty of deployment. Threats to cybersecurity have also been upgraded. On the regulatory side, on one hand, the existing financial regulatory framework is mainly designed for traditional business models and lacks effective regulatory tools for emerging businesses driven by AI technology. On the other hand, multinational financial institutions face compliance challenges arising from differences in regulatory standards across jurisdictions.

Beijing Business Daily reporter: Song Yitong

		Sina statement: This message is reposted from Sina’s partner media. Sina.com publishes this article for the purpose of disseminating more information, and does not imply agreement with its viewpoints or verification of its descriptions. The article content is for reference only and does not constitute investment advice. Investors act on their own responsibility and bear the risks accordingly.

Massive information and precise interpretation are available in the Sina Finance app

责任编辑：高佳