How AI Can Help Banks Decide Which Struggling Borrowers Deserve Help

Amid the pandemic, banks have “frozen” at least $150 billion of loans – equating to about 5 percent of their consumer portfolios. The question becomes: What happens when the freeze thaws and lenders try to get paid?

Unemployment, of course, has been skyrocketing, and tens of millions of individuals are having trouble making payments on bills. By the time they default, collecting on what’s owed may be a lost cause.

But enterprises can take a proactive approach to helping customers meet their financial obligations while maintaining their credit profiles at the same time. Identifying consumers who are at risk for missed payments or defaults are keys to that proactivity – and it’s critical to have the right data in hand.

Symend’s Hanif Joshaghani, co-founder & CEO, and Co-founder Tiffany Kaminsky told Karen Webster that the current economic environment exposes the limitations of risk scoring models used by companies across any number of verticals.

“Whether it’s a service or capital, when it’s a credit-based consumer product at huge scale – from a bank or a telco or a cable company – there’s a huge amount of decisioning and scorecards. It’s all about risk management,” said Joshaghani.

But these same firms are finding it impossible to keep pace with the breakneck speed of the pandemic, which has transformed the financial reality for tens of millions of people who have been thrown out of work.

Wrestling With Black Swans

Those black swan events, noted Joshaghani, happen more often than one might suppose, with a nod to the financial crisis last decade, and even California’s recent wildfires.

“The credit models cannot keep up with it – and as a result, you struggle to understand these consumers and how to decision at-risk individuals,” he said.

Leveraging digital engagement can help firms cement relationships with these past-due clients. The firm uses behavioral science and artificial intelligence (AI) to determine who has “intent to pay” and then craft payment plans.

Joshaghani said the challenge is for companies to avoid running into a wall where a huge number of customers just walk away from their debt and payment obligations.

As he put it, the platform helps financial institutions (FIs), utilities and others answer the question: How do firms bring their consumers along, step by step, and hold their hands to get them to make small commitments to pay back what is owed?

The company, which closed on a $52 million financing round earlier this month, said it has treated 10 million at-risk customers through the digital engagement platform to date, and will be on track to treat 100 million by the end of this year.

Symend’s Kaminsky explained that the firm’s white-label platform uses advanced analytics to derive dynamic insights into consumers’ situations and behaviors and deliver them to client firms.

“We’re using that AI and behavioral science to tailor a solution for each and every customer based on their preferences,” she said.

Kaminsky noted that the fluid situation of the pandemic requires dynamic customer sub-segmentation to create those personalized plans. By way of example, consumers who had jobs two months ago were low-risk at the time – and in the wake of the pandemic, they may have entered a payment deferral program. But she noted, too, that there must be education about what such deferral programs entail and what obligations continue to accrue.

And in an effort to tamp down massive debt that can become insurmountable, Kaminsky and Joshaghani said it is far more preferable to help consumers reduce debt bit by bit, with insight into what’s affordable – and where the companies recover something of what is owed.

Kaminsky noted that the firm has been measuring how consumers prioritize the services that they don’t want to lose, such as internet or phone, over, say, credit cards – and tailoring the engagement efforts (consider them nudges) accordingly.

In the end, the engagement platform can help create brand loyalty. And in looking ahead, with the data and learning gleaned from consumer behaviors with the help of advanced technology, the Symend executives said it will be possible to help larger credit software companies create better models that gauge risk before individuals ever become customers.

Joshaghani offered up an analogy: Where Symend has been in the Nyquil and aspirin business, helping clients address aches and maladies, eventually it can be in the vitamin business, warding off ailments before they strike.

“There is not going to be a silver bullet that solves this for everyone,” he said. “We’re talking about incremental improvements, which at a portfolio level will aggregate to having a meaningful positive outcome.”