China's banking sector is accelerating its push into artificial intelligence, weaving the technology into key areas from customer acquisition and decisionmaking to operations and risk control, as it races to build smarter, more responsive service models.

The effort is being reinforced at the policy level, with the country's central bank calling for the active yet prudent advancement of AI applications in the financial sector in a secure and orderly manner, unlocking the momentum for digital and intelligent development.

As the first digital bank in Jiangsu province, Jiangsu Su Merchants Bank said that more than 90 AI applications powered by self-developed and externally introduced large models have already been implemented across its business scenarios.

In smart lending, the use of multimodal large models has raised the overall accuracy rate for identifying nonstandard credit documents to over 97 percent, while improving end-to-end approval efficiency by more than 20 percent. In risk management, a large model-based decisionmaking system now delivers millisecond-level responses across the full loan lifecycle — from pre-loan to post-loan — boosting the accuracy of fraud risk labeling by 35 percent.

Currently, the bank's large-model AI technologies have been deeply integrated into more than 60 business scenarios, providing more precise and efficient financial services to over 4,000 technology enterprises and more than 17 million retail customers. In June 2024, MYbank and the College of Computer Science and Technology at Zhejiang University jointly established a laboratory focused on intelligent micro and small business finance, with an emphasis on the deep application of large models in banking.

Generative models are used to learn risk distribution patterns from vast datasets of small businesses, extracting more discriminative risk features. On this basis, discriminative models make credit decisions, while human experts continuously provide feedback to calibrate model performance, enabling ongoing iteration and optimization to more accurately map the credit risk profiles of small and micro enterprises.

In one vertical application developed with partners, MYbank has built a collaborative intelligent support platform that has enabled it to provide efficient and convenient credit services to more than 68 million small business operators. In credit approval, the consistency rate between AI-driven decisions and human judgments has risen from 39 percent to 90 percent, significantly improving efficiency while freeing up human resources.

Bank of Beijing, in collaboration with the Mentougou district government, recently launched its intelligent computing platform, marking a key step in advancing its "All in AI" strategy.

The platform has established a full-chain system from computing resources to front-end applications, embedding AI capabilities across core financial scenarios such as marketing, risk control, operations and lending. It enables unified management of enterprise-level AI computing resources, improving operational efficiency and reducing overall hardware costs.

At the same time, the bank has maintained a cautious approach to technological development, emphasizing the need to address risks such as AI hallucination, knowledge base quality and data security. It stressed that AI currently serves as an assistive tool and cannot fully replace human judgment.

A report released by PwC on March 17 shows that in the banking, insurance and asset management sectors, the majority of respondents view AI as a core engine of strategic transformation rather than merely a tool for improving efficiency.

Wang Jianping, a partner at PwC China, said surveyed financial institutions have already achieved initial returns of 10 to 15 percent on their AI investments. He added that while focusing on short-term gains, these institutions are placing greater emphasis on AI's long-term value in enhancing market position, expanding strategic development opportunities and identifying new growth avenues.

Xiao Yuanqi, vice-minister of the National Financial Regulatory Administration, said at the 2025 Bund Summit in Shanghai that AI applications in finance are mainly focused on process optimization and external services, spanning three key areas: the intelligentization of mid — and back-office operations, customer engagement and financial product delivery. Xiao said AI delivers dual benefits — helping financial institutions reduce costs and improve efficiency, while enabling more customized and targeted products and services.

He noted that AI applications in the financial sector are still at an early stage and remain primarily supportive in nature, unable to replace human decisionmaking.

Source: China Daily

DISCLAIMER: THIS ARTICLE IS FROM THE INTERNET AND DOES NOT REPRESENT THE OPINIONS OF 鹏盛资本PENGSCPA. 鹏盛资本PENGSCPA DO NOT GUARANTEE THE ACCURACY OR COMPLETENESS OF THE ARTICLE, WHICH IS FOR YOUR REFERENCE ONLY. IF ANYONE SUFFERS DIRECT, INDIRECT OR RELATED LOSSES DUE TO THE USE OF THE MATERIALS IN THIS ARTICLE, 鹏盛资本PENGSCPA WILL NOT BE LIABLE FOR SUCH LOSSES.