overview

AI Applications • May 9, 2023

Modernizing Collection Management with Artificial Intelligence

Collecting dues from customers is a tricky business, be it credit card, auto loan, mortgage, or any other loan that needs to be repaid. A high degree of customer empathy is a must for collection agents while interacting with consumers. I worked in a Mortgage servicing company as a ‘Home Retention Consultant’ and my job was to save their homes (or change non-performing loans to performing ones). The challenges included lots of relentless calls and a tremendous amount of persuasion, sometimes being cold and heartless as there were targets to achieve. This created a negative customer experience as they were bombarded with constant calls & reminders. Financial institutes tend to set up an in-house collection management department or outsource it to a third-party vendor. In both cases, it is expensive to operate and involves regulatory and other types of risk.

These problems got exacerbated by financial difficulties faced due to the pandemic. The world has not seen a more uncertain time after the recession of 2008. 14.7 million credit cardholders in the USA have defaulted in 2020, auto loans have around 3 million accounts that are in default. For mortgage, the number is at 2.7 million as per a report from Business Insider. With predictions of the high unemployment rate persisting for a prolonged period, credit card and loan defaults will also continue to rise.

The defaulting customers broadly fall into three categories according to their risk profile.

  • Low Risk: These customers have unwillingly defaulted their payment and wish to make the payment as soon as they know about it. This customer base is generally the ones who have forgotten their due date (These are customers who are 1 to 30 days past due)
  • Medium Risk: These are customers who are between 30 to 60 days past due. These customers may be facing a temporary issue, but they can fall either way to low risk or high-risk categories
  • High Risk: Customers who may not be able to pay their dues and are 60 to 90 days due. High-risk customers are struggling financially and will need a repayment plan

Apart from this categorization, some customer accounts are considered bad debts for banks and are generally outsourced to third-party debt collectors. Upon collection, a percentage of the collection amount goes to the collection agency.

However, collections contact center setup is expensive as it needs IT infrastructure and skilled resources. There are several regulatory laws that banks need to comply with while handling financial data, which makes outsourcing some of these jobs difficult. US FDCPA (fair debt collection practices act) and other applicable US state laws highly regulate the business. There has to be a statement given at the start of every collection call: “This is an attempt to collect a debt, any information obtained will be used for the same”. Missing this part can lead to a compliance violation. Additionally, collection agents have to be mindful of the time zones that they are calling as there are restrictions for the same.

Leveraging AI in Debt Collection and Management

A large number of difficulties in the debt collection and management space can be tackled with AI/ML and big data superpowers. We can leverage Doc AI to analyze the customer’s income documents by looking at their pay stub or other income documents. Customer’s debt to income ratio, property information, demographic data can be used to recommend a suitable repayment plan. The customer’s geolocation data from the call can help confirm if the customer still resides at the address mentioned in the bank records.

Moreover, information collected from customer credit history, social media, and third-party data sources can help predict customer’s delinquency risk, allowing creditors to take proactive steps to avoid bad loans. We can leverage Contact Center AI to send hyper-personalized automated messages to the customer through a medium best suited for the customer. It helps create a seamless yet effective communication channel while minimizing the agent’s involvement in the process. Automated calls, texts, or emails can be replaced as reminders of their payments for low-risk customers. Less dependency on human agents also means better adherence to the customer’s timezone. Chatbots can also set up automatic payments for low and medium-risk customers and ensure fund collection within time while updating the account simultaneously. Apart from this, we can offer a series of repayment plans for high-risk customers to avoid further default.

Adopting a cloud-native, intelligent debt collection system will not only reduce operational expenses but a model-based decision-making system will also minimize bias. It will reduce the chances of regulatory and compliance incidents as well. Most importantly, the pay-as-you-go scalable cloud deployment model will allow tier 2 and tier 3 banks to have their intelligent contact center setup rather than depending upon a third-party BPO or a collection agency.

AI/ML is revolutionizing the entire Banking and Financial Services industry and has opened doorways of new possibilities. It allows easy assessment of financial documents, analyzes customer data, builds effective communication channels, and facilitates personalized repayment plans to reduce default while minimizing regulatory and compliance risks. Thus modernizing the debt collection business is only a natural extension of that transformation.

Aniruddha Pisharody

Author

Aniruddha Pisharody

Top Trending Blogs

Blog
Banking & Financial Services

Quantiphi’s Journey: Pioneering AI Excellence and Earning...

20 September 2024

Start Your Next Gen AI Journey Today

Discover how Quantiphi’s AI-powered solutions can transform your business. Fill out the form, and we’ll help you explore tailored AI strategies to unlock new opportunities for growth.

Thank you for reaching out to us!

Our experts will be in touch with you shortly.

In the meantime, explore our insightful blogs and case studies.

Something went wrong!

Please try it again.

Share