Headquartered in Shanghai, Credit X is blazing a path in the financial services sector, exporting AI technology and machine learning to provide credit rating services for individuals previously overlooked in the China market.
Established in December 2015, the company has already completed Series B fundraising with total capital of 100 million RMB (US$15.4 million). Company founder, Zhu Mingjie, is a machine learning expert and worked previously with Microsoft Research, Yahoo, eBay, and Chinese travel agency, Ctrip.
Despite a successful career track and an enviable salary, Zhu was inspired to pursue his own business in a sector which he described as “low hanging fruit” due to an incomplete credit rating system which doesn’t account for individuals with low net-worth.
"Data is the core of financial services and essentially represents risk and credit, but we haven’t done a good enough job in this area," says Zhu. He set out to solve this problem by recruiting former colleagues to develop better credit risk models, soliciting experts with expertise in search, image recognition, and e-commerce referral systems.
Machine learning improves data quality
At present, most global financial institutions use the FICO credit rating system developed in the United States, calculating credit ratings through loan records, income, age, real estate, and auto ownership. Taiwan has similarly adopted this standard, but it isn’t applicable in China as only 15% of Chinese citizens have a credit record because most don’t have sufficient data to compile such a report, according to Zhu.
Banks, for their part, have limited manpower to compile information on low net-worth individuals because of the prohibitive cost of human resources. Thus, they are unable to service a group who have little choice but to take out high interest loans through non-bank networks.
The emergence of machine learning has the potential of turning "weak data" or incidental personal information into something useful in calculating credit scores. For example, mobile payments and other online transactions are now ubiquitous in China, and generally involve a large amount of data as they utilize mobile phones and reveal data such as work location that can help determine if an individual is a good credit risk.
Another example of how micro-finance institutions determine credit risk is analyzing a user's WeChat circle to determine how likely the user is to ask for money. Based on the user's habits when using the app, it is possible to determine whether the person is genuine or using an assumed identity created for fraudulent purposes.
Zhu explains that if a person's cell phone utilizes dozens of lending apps and is in use from 9 a.m. to 6 p.m., this could be a case of professional fraud. On the other hand, if the individual’s cell phone mainly accesses news apps and games it is more likely to belong to a “normal” person.
Zhu points out that such information cannot be evaluated by FICO’s simple mathematical credit rating model as it does not take account of weak, unstructured data. CreditX aims to deploy machine learning in this field to provide this expertise.
Not only useful for banks but also new fintech SaaS services
Positioned as a neutral, independent agency, CreditX envisions three categories of clients. The first includes traditional top-10 banks like current clients, China Minsheng Bank and China Merchants Bank. Zhu says strict regulations prevent banks from allowing their data to leave their premises so CreditX deploys staff to work inside the bank to provide analysis and examine user app habits to provide more complete data.
Another client segment includes internet-based financial institutions – typically consumer-oriented fintech pioneers. Zhu says these new start-ups are typically not initiated by banks nor technology companies, meaning they lack access to credit data as well as technology, making them potential SaaS service customers.
This essentially means that risk assessment, data analysis, and credit calculations, are all done on the CreditX cloud. Such customers simply create preset guidelines related to credit score rating and maximum pre-approved loan amounts.
Finally, there’s companies like Sesame Credit, a social credit scoring system developed by Ant Financial Services that is based on data provided by Ant parent Alibaba, or ZhongAn Online, a digital insurer, which have strong technical backgrounds but may be a competitor to many potential clients. Zhu believes staying out of banking services prevents a conflict of interest with banks and financial institutions.
Original News is from Business Next.