Behavioral AI is a Solution to Post-COVID Economical Changes, Especially Debt
Behavioral Signals, an emotion AI company specializing in behavioral signal processing, hosted a webinar titled “2 Ways Behavioral AI Can Help You Improve Your Debt Repayment Efforts” on Thursday, May 7th, 2020. Aris Karanikas, head of business development, and Dr. Nassos Katsamanis, vice president of engineering, co-hosted the event and Vicki Kolovou, head of marketing, acted as the moderator. Karanikas, also an economist, started by addressing how the current COVID-19 pandemic is going to affect the economy on both a macro and micro level. Dr. Katsamanis later discussed how institutions such as banks can apply behavioral AI technology to combat these economic changes.
Karanikas explained how one of the key pieces of data that we must look for in financial institutions is the amount of Non-Performing Loans, also known as NPLs. As more and more loans become non-performing (90 days past due or non-accruing interest) in an economic downturn, the economy as a whole suffers. Banks have to shift their focus to recover as much of these as they can while they set funds aside to provide for bad debt, funds that would otherwise be available for new loans. Thus, financing growth becomes harder.
Since the start of the COVID-19 pandemic, and as the United States and other countries around the world have gone into lockdown, we have been seeing staggering unemployment numbers. Even with the $1,200 stimulus checks sent out in the United States, many people are already struggling to pay back loans, while the stock market rises and falls on a daily basis, He pulled data to show that we won’t see the full scope of this recession until later in 2020, displaying predictions from the International Monetary Fund (IMF) where they expect the economic contraction to reach 5.9% in the US for 2020 and 7.1% in the EU. He compared these to the last recession in 2008 to put into perspective what we’re expected to face in the coming months.
According to the data, as an effect of this 2008 recession, NPLs in the United States grew from 1.01% in 2007 to roughly 5.3% by 2010. In the European Union, NPLs were at about 3% in 2008, and they more than doubled by 2012 when they reached 7.5%. It took between six to eight years for these ratios to recede back to their average values.
There is a set of challenges that financial institutions around the globe face while trying to collect debt. This sudden resurgence of debt repayment traffic has to be addressed with, at least in the short term, fewer resources, as they struggle to get their contact centers to full, remote operation.
But what if it didn’t need to be such a difficult process?
Behavioral Signals’ emotional and behavioral AI technology helps the process of debt repayment and restructuring, combining the benefits of optimizing performance, while maintaining personalized communication and customer satisfaction. Apart from knowing whether or not the person on the other end of the line is going to follow through with repayment, sometimes tensions can run high in conversations between the collector and the debtor.
Dr. Katsamanis continued the presentation and discussed how their machine learning technology is helping companies collect at a much higher rate. Their AI can seamlessly analyze conversations and build an extremely accurate profile of the debtor. The profile creates a matrix with a variety of categories such as a person’s confidence, aggressiveness, politeness, and more. This technology is all about analyzing how a person says something rather than what they say. By building these profiles on both debtors and collectors, organizations can better pair the two and increase the amount of debt collected.
In addition, Behavioral Signals has proven the power of leveraging emotionally intelligent data during a test they ran in the EU with a financial institution with roughly 200 agents. They chose 15 agents from the call center to form a challenger team, using this behavioral AI to analyze calls, and the test ran for a little over three months. The goal was to see if agents who used this form of machine learning could improve their amount of collections, and that’s exactly what happened.
Karanikas explained the results, and they were clear. During this time, the 15 agents connected with 4,900 debtors, and they were able to increase debt restructuring applications by over 20%. Not only that, but based on the data, they were much more efficient. Thanks to better pairings between agents and debtors and the ability to build better rapport, they reduced the number of calls by almost 8% during this time frame.
So, how did this affect the bottom line?
During the three-month testing period,15 agents increased active debt restructuring applications by 20%, which extrapolates $7.5M of additional restructured debt per year or $1.5M upside per agent per year. For a company with 200 agents, that’s an additional $300 million each year.
The webinar ended with an open Q&A where attendees asked to discuss the additional benefits of Behavioral Signals’ technology. The key usage is to match the right agents with the right debtors, but this technology is also great for training. Managers and agents can go over the data after calls to see where agents can improve with their style and tone to increase customer satisfaction as well. One of the other key benefits that they explained is that the conversations aren’t transcribed, so there are complete confidentiality and safeguards to protect sensitive information.
Although we’re living in uncertain times, Behavioral Signals is helping financial institutions around the world ensure that we come out of this pandemic with a stable economy. Their goal is
not only to help with the process of increasing debt collections but also to ensure that customers are satisfied every step of the way.