Standard Chartered Bank uses machine learning to analyze borrower behavior

Even though the BFSI industry is one of the early adopters of emerging technology, it is also the most vulnerable sector.

“An ability to distinguish between borrowers who are in default due to business externalities and potential willful defaulters is a serious concern for financial services providers,” says Zuzar Tinwalla, CIO at Standard Chartered Bank India.

Can financial institutions leverage technology to analyze borrower behavior?

There is no dearth of customer data available to the bank. But the problem lies in the fact that the data available is huge, scattered and often ambiguous, he points out.

Machine learning can generate meaningful insights, but if you feed data that doesn’t make any sense to the algorithm, that is exactly what you will get in return. To tackle this issue, Standard Chartered Bank India implemented a solution using algorithms to scan structured and unstructured data from internal and external sources.

The devil lies in data?

According to Tinwalla, executing the solution requires a combination of data integration (with the bank’s internal data), natural language processing (unstructured to structured semantic modelling), and a rules engine to asses risk and provide visualization for the end user.

The unstructured external data gathered from over 30 public sources is used to identify risks on entities of interest. For example, fire in a factory, court cases, labor disputes, or cancelled contracts, Tinwalla explains.

“We first built a prototype for the India Commercial Banking business and we are now developing and integrating the system with the bank’s infrastructure and core processing systems so that the solution can be used in other countries,” he points out.

Identifying risk                                                                                                                        

The solution deploys advanced analytics and visualization techniques to provide users intelligent and timely insights for better decision making. “The approach has helped us in acquiring crucial information and deep insights to enable oversight on both capital expenditure and operational expenditure of borrowers,” he says.

It provides early identification of risk, a 360-degree access to internal and external data, credit-monitoring and also ensures compliance with regulatory guidelines, adds Zuzar Tinwalla.