MYbank’s Jiang Hao: “The adoption of AI in banking is irreversible”

Ant Group, the parent company of AliPay, established MYbank in 2015 to address the financing needs of small businesses. By the end of 2022, more than 50 million owners of small and medium-sized enterprises (SMEs) had used the bank’s digital credit services to support their operations. Approximately 80% of these customers had previously been unable to secure loans.

The bank reaches out to the most excluded SME customers—those who borrow an average of RMB 4,500 ($635) for their firms—using its mobile app and cloud-based infrastructure. It pioneered the ‘310’ financing concept, where business loans take three minutes for application, one second for approval, with zero human intervention. By 2022, MYbank had earned $2.3 billion in revenue from SMEs, with a return on equity of 19.3%, higher than most traditional Chinese banks. Its total assets of $64 billion in 2022 trail only WeBank among Chinese digital-only banks.

President and executive director Jiang Hao said: “MYbank primarily serves MSME [micro SMEs] customers and e-merchants, and technologies are essential to the services we offer, hence our dedication to tech.”

The bank employs three solutions to meet their clients’ diverse needs: the Tomtit System, the Goose System, and the Lark System, that serve rural financing, supply chain finance, and loan risk management, respectively. Among these, Lark was showcased at the Singapore Fintech Festival.

The Tomtit system focuses on serving and incorporating rural consumers by including data obtained via high-accuracy remote sensors on satellites into their specialised credit underwriting algorithms, and then disbursing micro business loans in minutes.

Around anchor firms, the Goose System provides supply chain financing to MSME suppliers and dealers. MYbank manages to significantly boost the acceptance rate for MSME applicants by leveraging technology such as knowledge maps and deduction algorithms, as well as through collaborations with over 1,000 anchor firms.

While lending to the underbanked may raise some repayment concerns, MYbank developed the Lark system that uses artificial intelligence (AI) to monitor the entire lifecycle of analysis, approval, and post-disbursement to improve risk management that, according to the bank, is not only faster but also more accurate.

Jiang Hao remarked: “MYbank’s journey as a digital-only bank has more insights to offer to banking peers worldwide, not only in inclusive finance businesses, but also in technology solutions.”

Ongoing investment in technology enabled the bank to keep its gross non-performing loans ratio stable at 1.94% in 2022 while allowing millions of customers to borrow money. After using conversational AI, the bank observed a dramatic boost in customer care, with 90% of all problems resolved through its app and customer support hotlines.

As China’s first full cloud-native financial institution, MYbank has gained extensive experience in distributed architecture. It will be able to construct a fully cloud-native distributed architecture of SofaStack and OceanBase by leveraging the group synergy of Ant and Alibaba. With more elastic cloud-based business units and architecture, the bank improved scalability, allowing it to deal with concurrency scenarios such as e-commerce shopping events Double 11, as well as natural catastrophe readiness.

MYbank can automatically switch and allot all core transaction traffic and resources in any data centre, with over 10,000 physical servers spread over five data centres in three locations, all connected on OceanBase. As a result, MYbank’s design delivers a more flexible method of routing data traffic across the country, allowing the bank to use its resources more efficiently. Unlike peers, MYbank has not seen a single occurrence of fundamental system failure since its inception in 2015.

MYbank turns unstructured data into customer risk and requirements profiles in order to provide tailored services and product suggestions. Customer profiling, labelling, and analysis have been widely used in banking, however structured data-based profiling, such as card rating and risk rating, has proven insufficient to better understand clients. The bank now uses large amounts of unstructured data gathered from diverse channels such as instant messaging and social media in its data composition.