Personal credit issuance reaches 9 trillion yuan; Construction Bank executives detail AI applications

Ask AI · How does China Construction Bank use AI to enable customer managers to serve more than 20,000 customers?

China Economic News reporter Yang Jingxin, Beijing report

On March 27, China Construction Bank held its 2025 performance briefing. Unlike in the past, the bank’s artificial intelligence applications were highlighted as a key topic.

The bank’s annual report data show that by the end of 2025, its operating income reached 7610.49 billion yuan, up 1.88% year over year; net fee and commission income was 1103.07 billion yuan, up 5.13%; net profit was 3397.9 billion yuan, up 1.04%; and net profit attributable to shareholders was 3389.06 billion yuan, up 0.99%. The non-performing loan ratio was 1.31%, down 0.03 percentage points year over year; and the allowance coverage ratio was 233.15%, unchanged from the same period in 2024.

A reporter from China Business Journal noted that behind this impressive “scorecard” of China Construction Bank is a close relationship with the bank’s vigorous expansion of its retail business. At present, the number of the bank’s individual customers has exceeded 7.85 billion, and its credit disbursement volume has even reached 9 trillion yuan. Meanwhile, many retail businesses, such as personal loan business, wealth management business, and private banking business, all require “tailor-made” solutions, and artificial intelligence plays a key role in this.

Consolidating advantages in retail business

In the view of China Construction Bank President Zhang Yi, the bank’s 2025 “double-digit growth in loans in key areas” effectively supported the development of the real economy and practiced the financial “five major initiatives.”

Data show that by the end of 2025, China Construction Bank’s technology loan balance was 5.25 trillion yuan, up 18.91%; green loan balance was 6 trillion yuan, up 20.54%; inclusive finance loan balance was 3.83 trillion yuan, up 12.37%; core industries loans for the digital economy had a balance of 8919.26 billion yuan, up 18.7%; and the pension custody scale was 2.06 trillion yuan, up 16.87%.

It is understood that China Construction Bank’s total loans and advances exceeded 27.7 trillion yuan in 2025. Of this, domestic corporate loans and advances totaled 15.69 trillion yuan, up 8.7% year over year. Among them, loans directed to the manufacturing sector were 3.52 trillion yuan, up 15.83%; and loans to strategic emerging industries were 3.52 trillion yuan, up 23.46%. At the same time, the bank also provided 1.32 trillion yuan of financing support to 191.7 thousand chain-link enterprises under the industrial chains of 6594 digital supply-chain core enterprises.

However, compared with these corporate businesses, Zhang Yi believes that the retail business has greater advantages in terms of capital utilization and return. He said clearly: “China Construction Bank’s domestic personal loan balance is 9 trillion yuan. The share of businesses with low capital occupation and high returns reaches 32.59%, ranking at a leading level in the industry.”

From the perspective of profit contribution, in 2025 China Construction Bank’s personal loan yield was still considerably higher than that of corporate loans. According to data disclosed in the bank’s 2025 annual report, China Construction Bank’s corporate loan yield in 2025 was 2.76%, while the personal loan yield was 3.15%. In addition, retail businesses such as wealth management and credit card business can also effectively drive growth in non-interest income.

In fact, the reporter found that this annual report features plenty of standout data for the retail business. For instance, the scale of the bank’s personal operating loans reached 1.32 trillion yuan, up 28.77%; and personal consumption loans were 6831.74 billion yuan, up 29.41%. Tang Shuoxuo, vice president of China Construction Bank, commented that the bank’s personal consumption loans have seen additional increases of more than 100 billion yuan for three consecutive years.

In addition, the total number of the bank’s individual customers reached 7.85 billion; managed financial assets of individual customers exceeded 23 trillion yuan; and the number of wealth management customers also achieved double-digit growth.

AI applications “6 major areas”

“In the process of serving customers, a bank customer manager’s brain can provide memory capacity to serve 20 customers, 30 customers, or even 100 customers. But with the ‘Bangdei Assistant’ system, he can manage more than 20,000 customers.” Lei Ming, vice president of China Construction Bank, said.

So-called the “Bangdei Assistant” system is an artificial intelligence application system developed by China Construction Bank for customer managers. It is applicable to both retail and inclusive finance businesses, and provides customer managers with assistance such as customer insights, product recommendations, financial analysis, and customer outreach.

China Construction Bank said that it is actively advancing digital-intelligent transformation, equipping individual customer managers with “digital doubles,” deeply integrating AI technology into customer managers’ workflows. Around core business scenarios such as demand insights and asset allocation, it builds multiple vertical intelligent agents. Through data-driven decision support and intelligent analysis, it helps customer managers accurately grasp customers’ potential needs, enabling market opportunities to be “seen more clearly”; by leveraging specialized product analysis capabilities, it helps customer managers provide differentiated financial services, ensuring asset planning is “more service-oriented”; and through workflow large-model automation processing, it shortens the business handling lag for customer managers, so that customer needs are “responded to faster.” With the widespread application of “Bangdei Assistant,” it has significantly improved the digital-intelligent service capability of the personal customer manager team, achieving a dual enhancement in service effectiveness and quality in serving customers, and driving high-quality development of the retail business with solid technological capabilities.

It is understood that China Construction Bank has currently built an operating closed loop of “insight—matching—delivery” for direct customer management in the personal segment, with more than 74 million customers served online. By systematically promoting the construction of AI application systems, AI large-model technology has been scaled to empower 398 application scenarios across the group.

Lei Ming said that, at present, China Construction Bank has already introduced the most advanced algorithm models in the AI field, including DeepSeek, Tongyi Qianwen, and others. It has established 400 product applications, comprehensively covering the six major areas of intelligent marketing, intelligent customer service, intelligent investment advisory, intelligent operations, intelligent management, and intelligent risk control, namely AI applications “6 major areas.”

It is understood that China Construction Bank has currently advanced “AI applications for the entire credit approval process” in credit granting business, building a credit approval model of “AI-assisted + expert decision-making,” using technological innovation to help achieve a dual improvement in the quality and efficiency of credit approval. This application is based on a professional knowledge base for credit approval covering public and private domain data. Leveraging the autonomous perception and planning capabilities of multimodal large models, it fully refines latent experience of senior approval experts and transforms it into explicit capability libraries such as thought chains. It summarizes and forms key core risk points for similar businesses and the bank’s differentiated credit granting strategies, ensuring smooth risk preference transmission throughout the entire process. Meanwhile, with the application, within 1 minute it can generate a draft of rating review opinions including five modules; it can automatically compare the compliance of comprehensive financing business with policy requirements and draft more than 10 pages of review opinions; the maximum generation ratio reaches 90%, and reviewers only need to make minor refinements before submission.

Data show that after “AI applications for the entire credit approval process,” China Construction Bank’s volume of AI-assisted approval business in 2025 increased, but the average net time required for approvals decreased by more than 30% against the trend.

(Editor: Zhang Manyou; Reviewer: He Shasha; Proofreader: Yan Yuxia)