Machine Learning Is a Game Changer for Compliance

There are countless reasons to go digital in collections in 2020. Many of those reasons have to  do with the end user: the borrower. A growing proportion of consumers and borrowers are tech savvy Millennials who expect SMS reminders and sleek web portals. They don’t like to be  contacted on the phone and have a reasonable expectation that new numbers calling them are  scammers. 

As collection strategies move to digital communications, lenders can present tailored,  personalized communications. True machine learning provides the opportunity to mine data and  decipher trends hidden within your consumer behavior. Recognizing these trends guides the  creation of intuitive self-service customer experiences that will cultivate customer loyalty and  satisfaction and ultimately improve collection rates.  

But borrowers are not the only ones who benefit with today’s machine learning in debt  collections. Used appropriately, machine learning can guide new levels of regulatory compliance  and audit-proof your operations. 

Augmented Intelligence 

Humans aren’t perfect. That much is obvious. The vast majority of compliance problems in debt  collection come from manual processing problems—agents saying things they shouldn’t,  accidentally calling too many times—or from poor automation. There is a lot of room to improve  and tweak the system using augmented decision-making from machine learning. 

Machine learning is at the heart of Katabat’s strategy engine. But unlike many of our competitors  who advocate turning over decision making entirely to AI, Katabat’s machine learning leverages the power of both human and artificial intelligence to create intuitive and flexible workflows. Everything Katabat does is founded in the expertise of our operational managers. They draw on  their extensive knowledge of collections regulations and best practices to draft proven and  actionable strategies for SMS, email, and phone that are a gold standard. This strategy and  template repository is available to our clients so they can choose communications that meet  their compliance requirements.   

With automated outreach in place, the machine learning does its job of chewing through  massive quantities of data about the borrowers and debts. It predicts optimum outcomes and  selects the best block of actions from among the human-written strategies to deliver a superior, personalized customer experience.  

Machine learning is not going to solve everything for you. And it shouldn’t. The technology is still  relatively new, and while computers can now play chess or Go, they can’t come up with great advice all on their own. (Witness the really terrible results of efforts to have chatbots give  medical advice.) Nor can they generate emails and letters that comply with all the relevant  regulations. Many experts agree that the best use case for machine learning is to augment  human agents’ performance and vice versa. The ML and human partnership is where value is  skyrocketing right now, but we’re the only ones who are really talking about this in collections. 

Transparent Decision-Making 

In artificial intelligence, people talk about “black boxes” when there is a serious lack of  transparency. With the way some systems are built, it’s hard to gain understanding of why the AI  came up with the answer it did. In some cases this is by design, when the programmers don’t  want to share proprietary secrets. In other cases, it’s because transparency simply wasn’t a  priority. Both of these situations can lead to the widely reported cases of biased AI you may have  heard about. 

At Katabat, we have made transparency a priority. We know it is essential to our customers. Not  only do they have to comply with FDCPA and TCPA, they need to be able to demonstrate that  compliance throughout their operations if there is an audit. We meet that need by providing a  detailed feature analysis. The report shows exactly what criteria the ML used to make a  recommendation and which data had the biggest impact on the choice. The only place I want to  hear about black boxes is in aircraft safety systems. 

Help Your Agents Help Your Customers Pay 

Machine learning amplifies the impact of your best agents. It informs their actions using  predictions based on more data than they could hope to assess, and supports them with  Katabat’s library of expert strategies.  Katabat solutions bring together the best of both human and artificial intelligence so you can reach your customers exactly how, when, and where they are most likely to respond. Satisfied  customers are more likely to not only complete the payment process but to pay more than the  average. In fact, our data shows a 38% increase in payments with our machine learning at work.


Ye cofounded Katabat in 2006 and continues to enjoy creating technology solutions to solve business problems. Ye’s deep experience in artificial intelligence, banking and internet technologies have and continue to shape Katabat’s product development and evolution. Prior to Katabat, Ye worked for Bridgeforce and Ensuredmail, Inc. Ye received a BS and Master’s degree in Electrical and Electronics Engineering from Sichuan University. He also has his Master’s and Ph.D. in Computer Science from the University of Delaware. Ye is an avid technologist and has published multiple conference and journal articles in the fields of computer vision, pattern recognition, and artificial intelligence.

More from

Get the Latest Accounts Receivable News, Insights and Best Practices from Katabat Delivered to Your Inbox

We respect your privacy and your inbox. We’ll only send you timely content that matters most. And you can always unsubscribe at any time.

Request a Demo

Are you interested in learning more about the Katabat platform? Schedule a demo today.