Machine Learning Driven Early Warning Systems for MSMEs and Banks in India

Recently India’s MSME minister – Nitin Gadkari recently said about appointing a committee to resolve any problem in the implementation of the Rs 3 lakh crore collateral-free loan scheme for MSMEs. The government estimates 15% of Rs 3 lakh crore MSME loans could turn NPA. India’s formal banking sector has been grappling with bad debts since years, and the broader economy is starting to face the repercussions.At present, the detection of fraud loans takes an unusually long time. Mostly Banks tend to report an account as fraud only when they exhaust the chances of further recovery. Among other things, delays in reporting of frauds also delays the alerting of other banks about the modus operandi through caution advices by RBI that may result in similar frauds being perpetrated elsewhere.

According to Professor Vikas Srivastava of IIM Lucknow says that from a credit risk perspective, in most countries where central banks are reasonably efficient, approximately 2.3 per cent of loans and 1.6 per cent of deposits on an average are at risk.Given India’s history of bad debts, moving beyond conventional credit models is no longer an option, but a necessity. There is an imminent need for large-scale adoption of technologies like AI, ML and Big Data in the credit scoring and underwriting process to facilitate better decision-making for lenders. After all, technology is the key to solving the critical challenges faced by the Indian banking and finance sector

Crediwatch  employ AI/ ML algorithms on alternative data points such as statutory payment statuses, litigations, media sentiment, GST invoice data, bank statements as well as traditional data points such as financial ratios, industry outlook etc. The company has completed the development of the enterprise version of its flag-ship product, Early Warning System. This product is compliant with the RBI framework and is based on a proprietary library of 190+ early warning signals. The system comes with a case management module to track each alert and manage post-alert actions from the respective portfolio manager.

With this at the backdrop, I propose an interaction/authored article with Ms.Meghna Suryakumar, Founder & CEO,Crediwatch who could talk about:

·         Why traditional Data and Alert-driven Early Warning Systems have failed

·         Addressing the trust deficit of banks and financial institutes by not only providing credit history but also alerting them with early warning signals

·         How Banks are working on predictive algorithms with Data intelligence companies to mitigate the risk of NPAs and bad debts

·          The impact of COVID-19 on  the repayment ability of the borrowers and demand for fresh loans.