Back in March 2020, when COVID-19 was just about making its presence known in the country, the Reserve Bank of India was quick to anticipate the fallout on the banking industry. It announced a moratorium on loan repayments for three months, and then further extended it till September 2020. This was done to help both borrowers and lenders tide over the economic uncertainty of the pandemic - those whose jobs and incomes were affected could be relieved from paying EMIs, and the banks would not be forced to classify those particular loans as non-performing assets (NPAs).
While this did provide some temporary relief, in the long run, it was like slapping a band aid on a gaping wound. These loans weren’t going anywhere, and once the moratorium lifted, those who were reeling under job losses and salary cuts were still in no position to repay. Add to it the complexity of paying additional interest and pending EMIs since the due amount wasn’t waived off, it was just put on hold for a bit.
Now, in 2021, things don’t seem to be getting much better. For instance, In April, 29 million auto-debit transactions worth Rs 22,000 crore conducted through the National Automated Clearing House (NACH) failed.
The RBI has recently predicted the possibility of a spike in bad loans to 13.5 percent of the total book by September 2021, from 7.5 per cent in September 2020. It’s evident that no matter how thoroughly lenders assess risks, some factors remain out of anyone’s control. And that’s where the need for a tech-driven collections system comes in.
Loan repayment is about a lot more than just reminding borrowers to pay. An effective collections system helps navigate a potential default by acting as a two way-link between the lender and borrower. This involves the adoption of technologies such as Artificial Intelligence (AI) and Machine Learning (ML) to analyze customer insights and identify relevant patterns.
A combination of the two can be leveraged to examine a customer’s digital interactions. Have their online transactions reduced? Have they not received their salary in the last couple of months? Variables like these could be indicators of a financial crisis.
Recognizing these signs in real-time is a much more accurate and holistic method of assessing possible risk than a static credit bureau score. ML models can rapidly incorporate new data as a borrower’s circumstances evolve, and thus determine the possibility of delinquency before it occurs. This information can help lenders focus on such at-risk accounts and take measures to prevent delinquency, such as bringing down the EMI amount by increasing the repayment tenure or doing other forms of restructuring.
A collection intelligence system that leverages real-time cash-flow insights to identify risky borrowers is the need of the hour for lenders. With ML models trained on billions of data points, and ability to segment borrowers based on their riskiness, such tools can prove to be a boon in the hands of a lender. Further, advanced tools can then recommend resource allocation accordingly and even suggest alternate eNACH dates for bounced payments.
If the pandemic has shown the financial industry anything, it’s that even the most thorough planning falls apart in the face of an unpredictable disaster. An AI and ML-based early warning system is the best bet for lenders who want to mitigate the risk of loan defaults and ensure a healthy portfolio, no matter the circumstance.
(Rajat Deshpande, Co-Founder, and CEO of FinBox--embedded finance platform that enables companies to solve lending value chain using low-code, tech stack platform)
