How explainable AI is improving 'buy now pay later' schemes

How we bank, where we bank and who we bank with are changing dramatically. These incredible shifts are being driven by increased customer expectations and the power of disruptive technology to meet them.If evidence were needed for such an assertion, the rise of Buy Now Pay Later (BNPL) would be a good place to start.

Consumers today are used to fast, seamless, personalized experiences from global platforms like Netflix or Amazon. It’s an experience that’s also expected from banking: an intuitive and embeddable journey through everyday transactions. BNPL provides such an experience.

Buy Now Pay Later is an alternative form of credit embedded at the point of sale. When customers choose a purchase, they can immediately benefit from a BNPL loan without being diverted to a separate financial services journey or provider. Decisioning is fast, hard credit checks are rare, and there is typically little to no interest charged.

While the idea of paying in instalments is not new, what’s groundbreaking is how today’s technologies – like open APIs, cloud and artificial intelligence – have created new levels of speed, scalability and seamless integration into consumer platforms.

This technology is driving rapid growth. Around 360 million people currently use BNPL, which is expected to triple to almost 900 million by 2027. The trajectory for this extrapolation is an expected market growth of 45% each year between 2021 and 2030. It would take the market size from $132 billion in 2021 to $850 billion in 2026.

For example, PayPal used Temenos’ cloud technology to launch their Buy Now Pay Later solution in multiple jurisdictions. The results were dramatic: 48 million BNPL loan applications have been processed in two years, with 750,000 on just one Black Friday alone.

The rapid rise of Buy Now Pay Later has coincided, however, with a worldwide cost of living crisis. The impact of rising inflation and energy prices has not eluded BNPL, with a recent study finding that 42% of consumers using the payment method have been late with at least one payment.

In a sense, the ease and convenience of Buy Now Pay Later create its challenges. Because customers can purchase items quickly and frictionlessly, they may do so more impulsively. Our inherent behavioural biases favour the concept of “buying now” and “paying later.” We prefer to focus on rewards in the present rather than on future implications. This inclination is known as “present bias.” Meanwhile, future spending obligations are often underestimated, a phenomenon known as “atypicality neglect.”

Further examination of these topics is perhaps best left to behavioural scientists. But what’s clear is that all of us in the BNPL value chain need to do more to help consumers make more responsible and informed spending choices.

As concerns for consumer protection have grown, regulatory scrutiny over Buy Now Pay Later is naturally following. Global regulators are increasingly looking at areas like credit checks, transparency and education.

And as many commentators, such as S&P Global, have noted, such probing provides an opening for banks to enter the market when they have already lost billions in annual revenue to pure-play Buy Now Pay Later providers.

Banks are used to regulatory scrutiny in a way that new providers are not. They’re also well capitalized when borrowing costs for new players are increasing. But perhaps the most significant strength banks have is trust from their customers. In fact, 50% of customers prefer BNPL provided by their bank.

What this relationship affords banks is a wealth of customer data, including expected monthly income, spending obligations, past transaction history and even job status.

Knowing more about their customers allows banks to offer different forms of BPNL. One example is “pay later” services, where banks can retrospectively give customers the ability to convert previous eligible transactions into a Buy Now Pay Later loan. Such retrospection would relieve those struggling with cash flow while allowing the bank to manage default risk by selecting those with pre-underwritten credit risk.

And when this customer data is harnessed by an AI engine that is transparent and explainable, such as Temenos’ patented XAI platform – rather than a “black box,” there is even greater potential to make a positive impact.

For instance, in a typical online purchase scenario leveraging our XAI-powered BNPL service, the customer is presented with the bank’s BNPL payment option at checkout. They are given the option of different Buy Now Pay Later “flavours” (instalment amount, frequency, duration, interest rate), which the bank has pre-determined. At this point, the customer will be presented with XAI-driven estimates regarding their future spending ability, such as expected net disposable income.

These estimates are coupled with easy-to-understand insights, in plain language, that explain the key drivers behind that calculated figure. Customers can also receive recommendations for budgeting adjustments they need to make to afford the various BNPL options better. The result is a “win-win” for all parties.

For banks, this positive use of customer data allows BNPL to be an additional revenue stream that builds loyalty and trust while controlling delinquency and compliance risks. In addition, by offering BNPL directly under their brand, banks can better engage with otherwise hard-to-reach, digital-native Gen Z customers who are major Buy Now Pay Later users. And for the customer, it maintains the frictionless BNPL experience while helping them make informed spending choices, countering inherent behavioural biases.

So, just as digital technology allowed the market to reach its current level of growth, explainable AI can unlock the sustainability of this growth in the future. It’s a great example of how we can improve banking with technology.