Lack of collateral, never borrowing from banks or difficulty proving income has made many young people, freelancers and small traders almost "invisible" to traditional credit scoring models.
However, the development of digital data and artificial intelligence is creating a different approach. Instead of just relying on borrowing history, banks can analyze cash flow, payment transactions, stability in financial behavior and interactions in the digital environment to evaluate customers.
This is also the direction being implemented by TPBank in lending activities on digital channels. According to the bank, technology can help identify customers who are able to repay debt but previously did not have sufficient conditions to access credit in the traditional way.
Finding potential customers from digital data
Sharing about lending services on digital channels at the meeting of the Banking Industry Steering Committee on June 19, Mr. Do Minh Phu - Chairman of the Board of Directors of TPBank - said that the shift of people to the digital environment is fundamentally changing financial needs.
If previously the credit system was mainly designed for large, medium and long-term loans such as buying houses and cars, now, millions of small transactions arising every day are forming a different demand for capital flows.
Customers not only need to borrow for major financial decisions but also have a need for small amounts of money, which are resolved quickly, serving business operations or daily spending.
People not only need to borrow to buy houses or buy cars. They need immediate finances to serve their daily lives. The financial system cannot stand outside the flow of the digital economy" - Mr. Do Minh Phu assessed.
According to data provided by TPBank at the meeting, the scale of e-commerce in Vietnam in 2025 reached about 430,000 billion VND, with more than 601,800 stores operating on digital platforms. At TPBank, about 99.5% of transactions are currently carried out through digital channels.
Changes in consumer and payment behavior also create a large amount of data, reflecting relatively clearly cash flow, affordability and financial stability of each individual.
From transaction data, invoice payment history, eKYC electronic identification information and the stability of digital interactions, banks can add more basis to assess the creditworthiness and repayment ability of customers.
According to the Chairman of TPBank, technology plays the role of a "lens" to help banks see potential customers who were left behind by old credit models.
This group may include young people who have never borrowed capital, freelancers who have difficulty proving periodic income, small traders doing business on digital platforms, or customers who do not have much data in the credit information system.
Although they do not have traditional credit records, these customers can still maintain stable cash flow, pay bills on time and demonstrate transparent financial behavior in the digital environment.
Digital lending is not just about putting records online
According to Mr. Do Minh Phu, digital lending should not be simply understood as transferring loan applications from paper to online.
This model requires banks to redesign the entire credit chain, from customer identification, data collection and analysis, risk assessment, approval, disbursement to loan management.
At TPBank, artificial intelligence is being applied in some stages such as multi-dimensional credit scoring, detecting fraud in real time, optimizing limits according to customer behavior and personalizing service usage experiences.
According to information from the bank, the entire loan journey, including eKYC identification, approval, disbursement and loan management, can be carried out online.
To date, TPBank said it has served more than 6 million customers who have been disbursed, of which more than 5 million customers operate regularly on the digital platform. On average, each customer incurs about 2.08 loans.
In the period 2023-2025, the bank's digital lending turnover is said to have doubled year by year. The bad debt ratio announced by TPBank is below 3% and continues to improve in the process of optimizing the governance model according to IFRS 9.
Despite promoting automation, the Chairman of TPBank believes that technology cannot completely replace the role of people in banking operations.
We can automate processes, but not automate customer understanding" - Mr. Do Minh Phu once shared at the 2026 TPBank General Meeting of Shareholders.
According to TPBank's AI-Native banking model orientation, AI is deeply integrated into operations and service provision. However, people still play a role in risk control, establishing ethical principles and being responsible for final decisions.
Expecting to narrow the credit gap
According to the Chairman of TPBank, in the future, digital lending may no longer be a separate product but become part of the financial infrastructure in the digital economy.
Shortening the process and expanding the ability to evaluate customers through data can help capital flow faster, while bringing banking services to population groups that are not fully served.
This is especially meaningful for small value loans, serving essential needs or supplementing short-term capital for business operations.
When these needs are not met by formal credit institutions, people may have to turn to informal lending forms, with high costs and potential risks.
Therefore, safe and transparent digital financial solutions provided by banks are expected to contribute to narrowing the credit gap and limiting forms of usury.
However, the process of expanding digital lending also poses requirements for data quality, information security, anti-fraud and bad debt control capacity.
From the experience of deploying at TPBank, Mr. Do Minh Phu proposed that the banking industry continue to promote interconnected data connection, strengthen risk warning information sharing and build a more flexible management mechanism for digital financial products.
According to him, a suitable legal corridor along with an effective data sharing mechanism will create conditions for banks to expand digital credit while still ensuring system safety.
In the context of consumer, payment and business activities increasingly shifting to the online environment, data and AI are becoming important tools for banks to better understand customers.
For TPBank, the goal of this process is not only to speed up loan approval, but also to bring official credit closer to those who were outside the service scope of the traditional banking model.
