South Africa's banking sector is in robust health. Yet with few indications that the interest rate tightening cycle is reaching its end or that global inflation is subsiding, their lending businesses face significant headwinds through 2023.
There is a fine line they need to walk between responsible growth of their lending books and minimising the risk of rising default rates in the months to come.
The Experian Composite Consumer Default Index (CDI), which measures rolling default behaviour of South African borrowers, shows a drastic deterioration for the last quarter of 2022. Consumer default rates increased across the board, including mortgages, credit cards, personal loans and vehicle loans. Experian warns that we can expect CDI to deteriorate further in the first quarter of 2023.
The financial sector faces further headwinds with the Financial Action Task Force (FATF) recently greylisting South Africa for not fully complying with international standards around the prevention of money laundering, terrorist financing and proliferation finance. Banks need to further strengthen Know Your Customer (KYC) and Anti Money Laundering (AML) checks to address this reality.
As banks grapple with this landscape, one way that they can significantly improve their books is through prudent use of artificial intelligence (AI) and data science. Dramatic advances in technology mean that banks can today rapidly deploy smart AI algorithms to support lending decisions with accurate, predictive intelligence.
Such algorithms enable banks to use hundreds, potentially even thousands, of data points to make smarter lending decisions. These data points allow banks to inform their lending decisions with data about hidden consumer behaviours, supplementing the data they have about consumers’ credit records, income and expenditure with new insights.
Supporting decisions throughout the consumer journey
Data points empower banks to improve the decisions they make throughout the customer journey. Starting from enabling banks to identify high quality prospects during the customer acquisition phase, through to helping them screen out customers who will not pass the credit scoring checks early in the application process. It can also help them identify risky clients in terms of AML/KYC regulations.
Given the high costs of customer acquisition and loan underwriting, this application of AI can have a dramatic impact on a bank’s efficiency. Reducing these costs can allow the bank to offer lower interest rates to its customers and still improve its profitability. For example, one bank that worked with CyborgIntell to use AI to predict which customers are likely to drop out of a digital application, reduced customer acquisition costs by 40%.
When it comes to making credit decisions, AI can help banks to triage applications far faster, more accurately and with less human intervention. The tech can offer a real-time score for the probability that the loan will go bad. Lenders can automatically approve low-risk applications and decline those that are high risk.
Humans can focus on the cases in the middle that require more nuanced evaluation or special conditions. As the system collects more data, the algorithm will become more intelligent through machine learning. Over time, the bank will be able to approve or decline higher proportions of loan applications in an automated process.
Better CX, faster turnaround times
By speeding up turnaround times for applications, an institution can dramatically improve the customer experience. Furthermore, leading banks will ensure that the decisions they make with the support of AI can be explained to the customer. This can help a customer understand what they need to do to qualify for a loan and defuse their unhappiness if they’re turned down.
Consider the example of a bank that used AI and machine learning to improve loan approval rates and minimise risk. It developed a best-in-class data science model that would allow it to explain lending decisions to customers in real time. It increased approval rates by 32% with no increase to the bad debt rate.
Perhaps the most important breakthrough for banks is that AI allows them to be genuinely forward-looking and agile in their lending decisions. Today, banks depend on information that might be outdated or backward-looking to predict customers’ willingness and ability to pay or the possibility that an existing customer is about to default.
But given the economic volatility we face, banks can’t afford to act on old information. If they move too slowly, they risk acting on potential default only once the bad loans are already in their books. With AI, they can deploy new models fast and respond to changing consumer behaviour before it starts showing in the form of lower or higher default rates.
AI and machine learning enable banks to assess massive quantities of consumer data, analyse consumer behaviour and track consumer journeys. This can help them prevent defaults, improving turnaround times for loan disbursals, and enhance the customer experience. It’s an invaluable edge for any lender in an unpredictable market.