Financial technology debate turns to predicting borrower stress before defaults

At a U.S.-India fintech dialogue in Mumbai, discussions included the potential role of advanced intelligence in assessing repayment capacity and supporting borrowers before they fall behind on loans.

Lenders have traditionally relied on credit scores, repayment records, income and existing liabilities to assess whether a borrower can repay a loan. Advances in artificial intelligence are creating scope to examine a wider set of financial indicators, including how a borrower's repayment capacity may change over time.

This raises a key question for the lending sector: can better forecasting help identify financial stress early enough for borrowers and lenders to respond?

The subject was raised by Krishnakumar K T, Chairman and Managing Director of Oleevia Grameen Credits, an RBI-registered non-banking financial company headquartered in Kerala, during the U.S.-India Super Intelligence Leadership Dialogue at the U.S. Consulate General in Mumbai.

The dialogue, organised in collaboration with the Payments Council of India (PCI) and the Fintech Convergence Council (FCC), brought together representatives from banking, digital payments, financial services and fintech to discuss emerging technologies and cooperation between India and the United States. The discussions covered artificial intelligence, financial inclusion, risk management and the development of India's fintech ecosystem.

Moving beyond historical credit data

Credit assessments largely draw on a borrower's financial history and present circumstances. These indicators help lenders evaluate applications, but they may not fully capture changes that could affect repayment over the coming months.

Speaking at the dialogue, Krishnakumar suggested that more advanced intelligence systems could help lenders assess these changes by combining financial behaviour analysis, cash-flow forecasts and relevant economic indicators.

"Artificial Intelligence is already helping banks understand a borrower's past and present. The next breakthrough should be Super Intelligence that helps us understand the borrower's financial future," he said.

The idea is to help lenders assess not only whether a person meets the conditions for a loan at the time of application, but also how changes in income, expenses or business activity could affect repayment capacity during the loan term.

Such systems would need to be tested for accuracy and used with appropriate safeguards. Their effectiveness would depend on the quality of the information available, the reliability of their forecasts and how lenders use the results.

From loan recovery to early support

The discussion also covered the role of technology after a loan has been disbursed. Current lending processes often include payment reminders, overdue alerts and collection measures. Predictive tools could potentially add an earlier layer of support by helping lenders identify signs of financial strain before a missed payment occurs.

"The purpose of Super Intelligence in financial services should not be limited to improving loan recovery. It should help us identify when a borrower is gradually moving towards financial distress and guide that person towards financial stability before a default occurs," Krishnakumar said.

Possible applications include budgeting guidance, cash-flow planning, repayment support and financial counselling. Depending on the circumstances, lenders could also explore suitable repayment adjustments or restructuring options where permitted.

Any such approach would require responsible use of financial data, clear consent where applicable and safeguards for customer privacy. Early-warning systems would also need to avoid treating every change in a borrower's financial activity as a sign of likely default.

Implications for small businesses and households

Predicting repayment difficulties may be particularly relevant for small businesses, self-employed people and households whose incomes can vary over time. A temporary fall in earnings or a change in business cash flow can affect their ability to meet scheduled payments.

For lenders, earlier identification of these changes could support more timely engagement with customers and improve risk assessment. Whether this translates into fewer defaults or lower non-performing assets would depend on how accurately the systems identify genuine financial stress and whether suitable support is available.

Krishnakumar argued that lenders should place greater emphasis on preventing avoidable defaults alongside their existing credit assessment and recovery processes. The broader objective, he said, should be to support borrower stability while managing the financial risks faced by lending institutions.

Balancing innovation with responsible lending

The discussion reflects a wider question facing financial services as AI capabilities develop: how can lenders use predictive tools to improve decisions without relying too heavily on automated assessments?

A forward-looking approach could bring together three stages of assessment: a borrower's past credit behaviour, their current financial position and an estimate of how their repayment capacity may change. The information could then inform timely support, subject to the lender's policies and applicable regulations.

However, forecasting financial behaviour remains complex. Income can change unexpectedly, economic conditions can shift, and predictions may be affected by incomplete or inaccurate data. Human oversight, transparency and appropriate data protections would remain important to any wider use of these systems.

The U.S.-India Super Intelligence Leadership Dialogue provided a forum for discussing the potential role of emerging technologies in financial services. Krishnakumar's contribution centred on using intelligence systems to look beyond historical credit records and consider how lenders might identify and respond to borrower stress earlier.