The deep integration of the digital economy and the real economy is accelerating the transformation and upgrading of traditional industries, and the financial industry is also continuously improving its digital capabilities. In this process, how to organically integrate digital capabilities with actual business and further deepen the integration of data and reality is becoming a focus of industry attention.
At the "Technology and Finance Integration Development -2023 Excellent Competitiveness Financial Summit" held by China Business Daily, Fei Haojun, Chief Algorithm Scientist of Qifu Technology, stated that in the process of integrating data and reality, digitalization should be avoided as a "castle in the air", and more attention should be paid to the problems that need to be solved in actual business, ultimately embedding digital capabilities into business processes to truly improve quality and efficiency.
Focus on practical business issues
Regarding the original intention of digital transformation, Fei Haojun stated that under the national digital wave, Qifu Technology takes it as its responsibility to assist in the digital upgrading of China's financial industry, and provides services to financial institutions guided by digital and artificial intelligence technology. In the service process, Qifu Technology always adheres to the principle of "what problems can be solved by digitization" and focuses on the business. From intelligent marketing to intelligent management, intelligent risk control, and various aspects of the entire financial business, Qifu Technology achieves services through digitization and artificial intelligence technology. At present, Qifu Technology has served over 155 financial institutions and over 200 million financial users.
A more important question that arises from this is how to organically integrate digitalization with actual business operations? Regarding this, Fei Haojun said, "The most important thing is not to treat digitalization as a 'castle in the air'. Specifically, do not do technology just because you want to do it. Always remember the starting point of all digitalization or artificial intelligence technologies - to understand and solve business problems. Starting from solving business problems, then why do we need to do this function, why do we need to do dataization? Finally, embed the entire digital capability into the business solution process. I think the ultimate organic combination is that one day we forget digitalization, and digitalization naturally occurs in every process from decision-making to progress of the entire business. This is a more perfect organic combination."
When it comes to the specific exploration of Qifu Technology in the digital process, especially in the particularly hot field of artificial intelligence in 2023, Fei Haojun said that in fact, Qifu Technology has been reserving large model capabilities since two years ago. After the explosion of ChatGPT, the ability of big models has come to the forefront, and Qifu Technology also launched its own financial industry big model in April 2023. Compared to "rolling" model parameters, Qifu Technology places more emphasis on the efficiency of the model and is the first to empower large-scale models at a scale of tens of billions of parameters. At present, within Qifu Technology, the large model has achieved good application effects in user management, seat management, intelligent marketing, and risk decision-making assistance.
Exporting value-added services to small and micro enterprises
In the process of digitization and serving the real economy, in addition to empowering financial institutions, fintech is also empowering other perspectives, such as small and micro enterprises as the "capillaries" supporting economic development, which is one of the key directions. In Fei Haojun's view, small and micro enterprises will face two major challenges: difficulty in financing and weak technological capabilities. In response to the problem of financing difficulties, Qifu Technology has helped more than 100 financial institutions serve over 9 million small and micro business owners. In the process, there are actually two difficulties:
On the one hand, the information of small and micro business owners is incomplete. In response to the issue of how to provide more credit enhancing information to financial institutions, Qifu Technology has developed customized OCR capabilities (optical character recognition, the ability to turn images into text) to help financial institutions recognize small and micro information. The accuracy of customized OCR capabilities is nearly 10% higher than that of traditional OCR.
On the other hand, due to the difficulty in identifying risk information of small and micro enterprises in traditional risk control, Qifu Technology has constructed a financial knowledge graph that connects the industry geographical attributes, investment and financing relationships, transaction relationships, and behaviors of small and micro enterprises, forming an extremely complex financial behavior network that can help financial institutions better understand small and micro enterprises.
Fei Haojun pointed out that the digital investment of small and micro enterprises is limited, so Qifu Technology has also systematically and systematically used its accumulated digital capabilities in serving financial institutions to help small and micro enterprise owners provide value-added services, such as bidding information, industry information warning for small and micro enterprises, and the acquisition of potential customers for small and micro enterprises.