Predictive Analytics in Financial Services
One way AI is used within the industry is through a predictive analytics strategy known as “next best action” or “next best offer,” says Jerry Silva, research vice president at IDC Financial Insights.
Banks have traditionally used this approach to anticipate things, such as when a customer might need a car loan, based on previous activity: the date of the last car purchase and the number of car payments, for example. Now, “there’s been a big shift,” Silva says. Using AI to analyze customer data is not only about selling someone a product but fixing a potential problem, such as deciding whether the bank should waive an overdraft fee for a loyal customer with many bank products.
“The next time you encounter the bank — you call the contact center, you’re at a branch — they now have an indication of, what should we do for this customer? How do we delight this customer instead of just trying to sell them?” Silva says.
Chatbots Continue to Grow in Financial Services
Banks are also increasing their use of chatbots, or computers that simulate human conversation. These have been growing in popularity over the past decade, but haven’t yet gained widespread use beyond the largest banks, despite being “the most visible consumer-facing use of AI” in the industry, according to Independent Community Bankers of America.
A December 2020 survey of senior executives at U.S.-based midsized financial institutions by Cornerstone Advisors showed just 8 percent of banks and 18 percent of credit unions were already using chatbots. However, that’s changing: Executives at 15 percent of banks and 18 percent of credit unions said they planned to deploy chatbots in 2021. These numbers are up from 3 percent at both banks and credit unions in 2018.
A big reason is the potential for cost savings derived from having machines, instead of people, manage ordinary customer transactions. Juniper Research, for example, estimates that the adoption of chatbots could save the healthcare, banking and retail sectors $11 billion annually by 2023, mainly by eliminating 2.5 billion hours of human work.
How AI Will Help Financial Institutions Collect Debt
Traditionally, banks have made only minimal efforts to collect payments on a nonperforming loan before selling the debt to a collection agency at a fraction of its face value. But some have turned to AI to help keep better track of these loans and reduce such money-losing sales.
The trend hasn’t taken off as quickly as Silva initially thought it would, he admits, but says there’s opportunity for growth, especially as the pandemic goes on.
“The bank would try to work with the customer — and again, here you need that customer data, you need to run the analytics,” he says. “You really want to approach collections differently, try to work with the customer, try to arrange a different payment schedule.”
The key is determining which customers are likely to pay up under the right schedule and pinpointing what that payment schedule should be.
“That’s certainly an area where the analytics, the AI would also help institutions in collecting more money just by working with the customer, versus just selling it off,” he says.