During the first half of this year, the pace of artificial intelligence deployment at China's major state-owned commercial banks has accelerated significantly, with steady progress made in technological investment, the scale of implemented use cases, and computing power support. This is according to insights gathered from the 2026 interim reports of the banks, which highlight a clear shift towards more sophisticated digital transformation strategies.
As a benchmark for industry digitalization, 工商银行 (ICBC) continues to lead in the overall application of AI technology. By the end of June, the bank had implemented more than 600 large model use cases, enhancing the precision of its services to the real economy, improving operational efficiency, and making risk prevention more proactive throughout the institution.
Similarly, 中国银行 (Bank of China) is accelerating the implementation of its "AI+" plan, with its large model ecosystem expanding rapidly. As of the end of June, its BOCAI large model had surpassed 32,000 active users, with over 2,200 intelligent assistants and more than 3,800 application scenarios in use, covering all business lines from front to back office. Meanwhile, 建设银行 (China Construction Bank) has launched over 600 AI application scenarios, and 交通银行 (Bank of Communications) has implemented more than 420 AI scenarios, empowering its core business operations.
In the AI era, the average daily consumption of tokens has become a key indicator for measuring the speed of banks' digital transformation. In this year's interim reports, some large banks have publicly disclosed their core token consumption data for the first time, showcasing the tangible results of industry-wide AI adoption. According to Niu Xinzhuang, Vice President and Chief Information Officer of 邮储银行 (Postal Savings Bank), the bank has implemented over 370 large model application scenarios, with more than 10 million daily model calls and over 30 billion daily input and output tokens. This supports the vision of providing "one smart branch for every customer externally, and one intelligent assistant for every position internally."
Beyond enhancing customer services and risk control systems, AI large models are now evolving into the "ultimate support" for employee empowerment and internal management optimization. In the personal finance sector, 工商银行 (ICBC) has deepened its human-machine collaborative marketing model. Its intelligent assistant for relationship managers helps with data retrieval, analysis, and proposal generation, enabling them to provide intelligent financial steward services to clients over 22 million times.
农业银行 (Agricultural Bank of China) is focusing on internal management and employee enablement with its self-developed enterprise-level intelligent agents and AI office system. According to President Wang Zhiheng, the bank has launched its own version of a "Lobster (ABC-Claw)" and created an enterprise-grade AI digital assistant named "一明". This assistant leverages a broad-intent recognition capability to enable collaboration across multiple intelligent agents in different business domains, automatically handling complex tasks such as customer marketing, credit processing, and intelligent office workflows.
建设银行 (China Construction Bank) has equipped every relationship manager with a dedicated "exclusive intelligent assistant" called "帮得" (Bangde). Through dialogues with this assistant, customer managers can gain insights into clients, configure products, and execute actions with a single click, covering nearly all aspects of their daily work. In the first half of the year, "帮得" saw an average of over 40,000 monthly active users and nearly 90 million total interactions. As Vice President Lei Ming noted, the tool has visibly improved customer managers' performance. Those using "帮得" effectively saw a 57 percentage point increase in new customer assets under management compared to the bank's average level.
邮储银行 (Postal Savings Bank) is focusing on high-value scenarios in retail customer service, corporate marketing support, internal operations, and employee office assistance. They have built the "智慧小邮" super customer service, an intelligent assistant for customer managers, and a super office agent. Among these, a retail marketing agent dynamically generates personalized campaigns, reaching 54 million high-potential postal savings customers.
Guided by the national "15th Five-Year Plan" strategy to "fully implement the 'AI+' action" and "seize the high ground of AI industrial application to empower all industries," several state-owned bank executives have outlined their next-phase AI development plans at their interim results briefings.
交通银行 (Bank of Communications) will deepen its AI applications in areas such as technology enhancement, employee empowerment, customer service, and risk management. As Vice President Qian Bin explained, "We aim to use AI to elevate the overall efficiency of our technology, moving business and technology from a mere 'handover of requirements' to 'co-creation'. We are building a digital workforce system with intelligent assistants, digital avatars, and digital experts. Through human-machine and multi-agent collaboration, we will drive job reinvention, process reengineering, and organizational upgrades."
"Technological exploration is just the starting point; system capabilities and application results are the real tests," stated Niu Xinzhuang of 邮储银行 (Postal Savings Bank). He emphasized the commitment to principles of self-reliance, practicality, and security, aiming to deeply embed AI into business logic and the entire product service process. This includes using autonomous, adaptive, and collaborative agents to achieve process reshaping, service integration, and experience enhancement, ultimately forging a new path for high-quality digital finance through a "general intelligent foundation plus diverse scenarios" application model.