Uday Chaudhari, senior chief technology officer for Indian operations at IT and consulting firm Synechron, said AI / ML enables banks to analyze the likelihood of credit events and proactively manage individual risks . “They allow banks to offer products and offers that reduce the risk of default on loans,” he told The Times Techies webinar last week.
Vijay Aggarwal, CTO of payments firm BharatPe, said AI / ML algorithms and data-driven decisions have helped fintechs keep APNs low. “Underwriting is where we have the most impact on data and ML compared to the traditional setup. We have a lot of historical and real-time trader data on our platform, and we can analyze it to assess the risk associated with each loan, what should be the amount and how long we should offer. The traditional method would be to review income tax returns, profit and loss statements, ”he said.
Anantha Sharma, chief technology officer at Synechron UK, said that in the financial services industry, an ML strategy must be implemented in a way that meets regulators’ expectations for explainability. ML models are often opaque – you cannot distinguish the contribution of each parameter to the predictions made. “Banks focus on statistical models rather than a deep learning model because deep learning models are by nature not easily explainable, although much has been done to make them as explainable as general statistical models, ”said Sharma.
Chaudhari said AI / ML is still in its early stages and big developments are yet to come. “Every AI / ML model has a learning curve. We haven’t crossed that curve yet, ”he said.