Six ways AI can help lenders refine debt collection processes

During these economically challenging times, South African consumers are under extraordinary pressure. Many have resorted to borrowing to cover their basic living expenses, with the Q2 2023 Consumer Default Index (CDIx) from Experian showing that South Africa’s 25 million credit active consumers account for just over R2 trillion of outstanding debt.

The Index, which tracks the rate of increase in consumer first payment default, has steadily increased over the past six quarters across consumers of all levels of affluence. Many lenders are responding to deteriorating economic conditions by tightening affordability criteria for loans, but that doesn’t address the challenges lenders face because of increasing default rates among existing borrowers.

For banks and other lenders, the efficiency of their debt collection process is thus emerging as a crucial factor in maintaining financial stability and profitability.

One major challenge of debt collection lies in ensuring that the cost of collections remains at a sustainable level, especially when dealing with high volumes of small debts. Collection is expensive, admin-heavy and time-consuming. What’s more, the longer a debt remains unpaid the more challenging it becomes to collect it. As debts age, they may become prescribed and thus uncollectible.

Lenders are thus looking for tools and technologies that enable them to collect debt more efficiently and effectively. Many banks are reported to be turning to artificial intelligence (AI) and machine learning (ML) to streamline their collection process, improve debt collection success rates and complement the expertise of human credit collection teams.

Six ways AI can lend a helping hand include:

1. AI-powered algorithms can analyse large datasets to identify patterns and trends in debtor behaviour. By segmenting debtors based on their payment history, financial status, and communication preferences, collectors can tailor their strategies for each group. This can vastly increase the chances of successful debt recovery.

2. AI can predict the likelihood of a debtor paying back their debt based on historical data, credit scores and debtor behaviour. By assessing the value of outstanding debts as well as the prospects of successful collection, collectors can prioritise collection of high-value debts from debtors that are most likely to cooperate.

3. AI can flag accounts that show early signs of delinquency, enabling collectors to intervene and decide whether to pursue a debt aggressively or adopt a more flexible approach, such as offering to change the repayment terms, depending on the debtor’s risk profile. ML-based risk scoring can assess the risk associated with each debt and debtor.

4. ML-powered chatbots and automated messaging systems can handle routine communications with debtors. These bots can send reminders, answer basic queries, and engage in interactive conversations, saving time for human collectors and ensuring periodic and consistent communication.

5. AI systems can analyse historical data to determine the best times and channels for contacting debtors and optimise contact strategies. By reaching out at the most opportune moments, collectors can increase the chances of debtor engagement and payment.

6. ML models can be used to scrutinise debtor financial data to recommend personalised payment plans that align with their capacity to repay. This tailored approach increases the likelihood of successful repayment and fosters a positive relationship between debtors and collectors.

The good thing about AI and ML is that the systems and the associated models can learn from past collection efforts and outcomes. As more data is collected and analysed, the ML system can refine its predictions and become more accurate over time. This enables banks to continuously improve their decisions and strategies for debt collection, enabling them to become effective and profitable over time.