At the 2025 Hong Kong FinTech Week, which opens on November 3rd, the deep integration of financial technology and artificial intelligence has once again become a global focus. This major event brings together prominent guests such as Hong Kong Special Administrative Region Chief Executive Li Jia-chao and Nobel Prize in Physics winner Geoffrey Hinton. Among them, Zhu Guang, CEO of Duxiaoman, delivered an important speech, detailing the innovative applications and disruptive impact of large language models (LLMs) in the financial services industry.

The "Customer-Centric" Revolution of Large Models: From Monthly Surveys to Second-Level Feedback

Zhu Guang pointed out that over the past year, large models have achieved a qualitative leap in global knowledge and reasoning capabilities, driving the transformation of the financial industry at an unprecedented speed. He emphasized that large models allow every employee of a company to be closer to customers, realizing the true concept of "customer-centric."

Real-Time Insights into Customer Needs: Traditionally, companies relied on time-consuming and laborious regular surveys to understand customer needs. Now, with the powerful data processing capabilities of large models, Duxiaoman analyzes approximately 60,000 to 70,000 customer interaction data daily, quickly extracting user feedback and needs, greatly accelerating the product iteration and innovation cycle.

Core Decision-Making Efficiency Soars: Credit Review Time Reduced to 30 Seconds

In the core decision-making scenarios of finance - credit approval, large models have even surpassed traditional machine learning models.

Improved Credit Review Efficiency: Zhu Guang revealed that Duxiaoman has applied large models to credit review work. By imitating the judgment logic of the best credit reviewers, AI assistants help employees reduce the credit review time from the original 10 minutes to an astonishing 30 seconds.

Risk Control is Guaranteed: At the same time that efficiency improves, risk control has also seen significant improvements, with risk indicators being reduced by more than 50%. This simultaneous improvement in efficiency and risk control not only significantly reduces labor costs but also greatly improves the response speed of customer service.

Realizing Inclusive Finance: High-Quality Services Are No Longer Exclusive to High-Net-Worth Clients

Zhu Guang emphasized that the application of large models is turning the concept of inclusive finance into an achievable reality.

Popularization of Financial Services: In the past, high-quality financial services were mainly concentrated among high-net-worth clients. Now, through the empowerment of intelligent advisors, ordinary users can easily obtain customized financial planning and suitable financial products.

Everyone Can Enjoy Quality Service: This means that whether it's a young person entering the workforce or a small shop owner running a business on the street, they can now easily and efficiently access financial services that meet their own needs.

Future Prospects of AI Assistants and Web3

Guest speaker Lin Chen, Deputy Vice Chancellor of the University of Hong Kong, shared his vision for the future of financial technology. He pointed out that in the upcoming Web3 era, everyone may have multiple personalized financial assistants, which can efficiently assist users in financial planning, credit, and investment decisions.

Lin Chen predicts that interactions between these AI assistants will become more frequent, achieving automated execution of financial services through smart contracts