Text | Yuxi
Let's start with two questions:
- What is the essence of micro and micro business credit?
- What is the essence of financial management for small and micro enterprises?
In the eyes of MYbank President Feng Liang, both of these issues are related to *forecasting*.
The former is the relationship between the future operating cash flow and liabilities of small and micro enterprises.
The latter is the opposite, which is the *prediction* of when small and micro enterprises will have money and when they will use it.
And *prediction* depends on AI.
Three Canadian economists, Ajay Agrawal, Joshua Gans and Avi Goldfarb, argue in their book Power and Prediction that AI is essentially a "prediction machine" from a business perspective.
Financial Prediction Machine
Prediction is a prerequisite for decision, and AI prediction can change the way people make decisions.
The improvement of forecasting ability has brought many changes to MYbank, and today's MYbank is becoming a financial forecasting machine.
Let's start with the following example.
Recently, with the release of the 2023 financial report, many media outlets have said that MYbank has "shrunk its balance sheet" and attributed it to insufficient credit demand and sluggish growth, but this is not the case. The real reason is that AI has improved MYbank's ability to predict users' deposit and withdrawal behaviors, which has correspondingly reduced liquidity reserves, and the corresponding asset scale has decreased. Isn't this kind of "balance sheet reduction" a connotative development path that the banking industry has in mind?
This is just one example of how forecasting leads to operational improvement.
In fact, MYbank has been doing this for 9 years, and if you go back to the predecessor of MYbank, Ali Micro Loan, the attempt to predict the repayment ability of customers actually started in 2010.
At that time, it was often said that "310" was "3 minutes to apply, 1 second to lend, and 0 manual intervention". At that time, this was possible precisely because Ali Microfinance was able to predict the repayment ability based on the user's data. Today, this kind of forecasting has been greatly iterated, and MYbank's credit forecasting can be based on a comprehensive analysis of the company's historical transaction data, industry trends, and market environment, which greatly improves the accuracy and efficiency of loan approval.
At present, AI has greatly improved MYbank's prediction capabilities in many aspects, such as credit approval, customer portrait analysis, industrial chain financial services, risk management, wealth management product matching, 24-hour financial services, enterprise actual controller identification, data security and privacy protection, etc., and has also been deeply integrated into MYbank's daily operations, enabling it to optimize its own risk management and operational efficiency while meeting the financial needs of small and micro enterprises such as lending and capital management.
Taking the distribution of bank wealth management products as an example, the current scale of MYbank's distribution has reached 610 billion, second only to China Merchants Bank in China, which also requires AI technology to predict the flow of customers' funds, so that MYbank can accurately match customers' capital allocation needs with the product supply of asset management institutions.
There are many more such examples.
When I asked Feng Liang what he thinks about AI, Feng Liang said that the current focus of AI is to replace the fields that have large user needs but require a lot of manpower.
He explained that if you use the traditional method, with the current business scale of MYbank, it needs at least 1-20,000 relationship managers and loan approval team, but MYbank currently has only 1,600 employees, more than half of whom are engaged in technology research and development.
In fact, a lot of the manual work that was originally required was to do prediction and judgment.
Before AI, banks were traditionally judged by rules, taking into account factors such as financial statement data, credit scores, mortgage or guarantee requirements, loan purposes, industry characteristics, and policy support. These rules are designed to prevent risks, but they also raise the bar and reduce efficiency, making financing for small and micro enterprises a long-term problem.
With AI, banks can be more accurate in predicting user behavior feedback, where predictions replace rules, and even decisions are made directly by AI.
AI Applications: At what stage is MYbank?
Although MYbank has done so much, in Feng Liang's view, looking back from the future, there is still a big gap in the middle, and there are still many areas where AI can be developed.
"Isn't it possible to go deeper?"
"It's too likely!"
So, what stage is MYbank at today?
In Power and Prediction, Canadian economists Ajay Agrawal, Joshua Gans, and Avi Goldfarb discuss how artificial intelligence (AI) is disruptively impacting economies and societies. According to the book's perspective, the application of AI usually goes through several stages: introduction, prediction, power, and transformation.
- Introduction: At this stage, AI technology is introduced into a specific industry or field and begins to replace some simple and repetitive tasks.
- Prediction: As technology evolves, AI is being used for more complex tasks, such as prediction and pattern recognition, which can help businesses better understand data and trends.
- Power: AI technology has become more powerful and is beginning to influence the decision-making process, providing in-depth insights and even making autonomous decisions in some cases.
- Transformation: Ultimately, AI technology will lead to fundamental changes in entire industries or fields, changing the competitive landscape and creating entirely new business models and markets.
Power and Prediction: The Disruptive Economics of Artificial Intelligence
In my opinion, MYbank's AI application is in a period of transition from the prediction stage to the decision-making stage.
The reason is as follows, MYbank is already using AI to analyze and process customer data, such as identifying, processing and processing images and data through the "Bailing System", and using AI technology to predict customers' cash flow and loan needs. At the same time, the application of AI in MYbank has begun to influence the decision-making process, such as providing support in credit approval, risk management, and wealth management product recommendation, making decision-making more accurate and efficient.
Further, according to Feng Liang, MYbank is exploring how to integrate AI technology into its business more deeply, such as re-evaluating the position and role of small and micro enterprises in the industrial chain from the perspective of industrial chain finance, and this evaluation can even go deep into the fourth or fifth level of the industrial chain (for details, see "Doing Industrial Chain Finance, MYbank Iterates on These Three Points"), which may eventually lead to fundamental changes in the bank's service model and product design.
Advantage: High-quality private domain data
So, the question then arises, can other banks that MYbank can do the same? How high is MYbank's moat in the matter of AI?
A machine learning expert, jbetker, posted a post on his blog in June 2023, saying that he has been working at OpenAI for a year, trained various models, configured various parameters, and found a pattern:
It's like no matter which student is smart and which student is stupid, as long as they all study hard, they will eventually learn the textbook used in school.
Feng Liang, president of MYbank, also holds the same view, saying that in the long run, it is difficult to produce a large degree of differentiation in the model. In his words, the real final distinguishing is the data, or the high-quality model training data, pre-training such data, especially in the professional field and private domain.
The author of this article, You Xi, and Feng Liang, President of MYbank
And that's where MYbank comes in.
MYbank is a bank that has grown on a digital platform, and after 9 years of data-based operation, it has a large amount of private domain data, which is real business data. According to the 2023 financial report, MYbank has provided loans to more than 53 million small and micro customers, and in the financial field, MYbank has served such a large number of customers.
Even if a customer has not yet taken a loan with MYbank, MYbank can enrich its data dimension through data such as Alibaba's e-commerce, Alipay's physical receipt and payment data, as well as the data uploaded by the users of the Bailing system themselves, and the data of satellite remote sensing of the Big system.
In Feng Liang's view, the next step for MYbank is how to combine the value of these data with the advantages of large models and AI, and refine and mine it.
Every new technology will reshuffle the industry landscape, and MYbank's AI potential will be further consolidated if its AI potential can be strengthened.
MYbank's application of AI has not only strengthened its competitive advantage, but also contributed value to the industry.
At present, 74% of MYbank's loan users have their data shared with other financial institutions on the central bank's credit report after completing their first loan. This is also critical for the rest of the country's financial system as a whole.
Therefore, in the process of writing the big article of digital finance in the whole industry, MYbank has written an important chapter of innovation.
Since its establishment on June 25, 2015, MYbank has gone through 9 years. As a private bank that prioritizes digital operation, MYbank has always positioned itself as a complement to traditional banks, focusing on inclusive finance and small and micro finance, and providing financial services to small and micro business owners, individual industrial and commercial households, and rural households. The loan amount of 70,000 yuan per household reflects its efforts and persistence in the field of inclusive small and micro financial services.
But from the perspective of the industry, I think what is more noteworthy is its development in the no-man's land, as well as the cognition and exploration of the industry. Ten years ago, MYbank used digital operations to make small and micro loans, creating a new business format, and this accumulation provided a solid foundation for them to transition from digital to intelligent.
Therefore, for practitioners, paying attention to MYbank means paying attention to the cutting-edge practices of China's banking industry in the field of AI and the latest progress of digital finance. For MYbank, it is also the significance of MYbank to adhere to the original intention of serving inclusive small and micro enterprises, continue to play the role of a supplement to the financial industry, and explore the future possibilities of China's banking industry.