Machine Learning in Payments - Applications & Opportunities During Crisis

The COVID-19 crisis has created a major shake-up for pretty much every person and industry on the planet in some way, shape or form, including the payments industry. With the requirement of keeping your distance from other people and a lot of bricks-and-mortar shops being forced to remain closed unless absolutely necessary, physical payments have been limited, retail and commerce have taken a heavy hit and B2B payments have been significantly reduced. The flow of money, in other words, has been reduced to a trickle.

And yet, all is not doom and gloom. Where established industries decline, alternatives will inevitably fill the void they leave behind. Artificial Intelligence (AI) has been doing just that, with machine learning enabling financial institutions, payment providers and digital businesses to join together and support their customers. In this article, we will take a look at exactly how that is happening.

How has COVID-19 impacted the payments industry?

The world has now been at the mercy of the coronavirus pandemic for over a year, with the numbers of infected and deceased already into shockingly high numbers and continuing to rise. As saddening as the human cost of the crisis has been, the long-term economic impact on the world is expected to be just as staggering.

According to McKinsey & Company, those industries hardest hit include airlines, tourism and retail, resulting in a massive reduction in cross-border transactions, disruptions for supply chains and B2B transactions and a significant decrease in retail and point-of-sale payment volumes. These results spell a negative revenue growth for the payments industry. After all, if people aren’t spending money or are spending it in different ways, how are payments firms going to make any money?

However, just as all of us are adapting to the “new normal”, so too is the financial sector, especially those able to innovate and adapt to changing circumstances. As the need for contactless transactions and online transactions increases, global payments firms are shifting their focus to digital methods of payment. These include:

  1. Using radio frequency identification (RFID) technology and virtual cards to enable contactless payments,
  2. Supporting remote-commerce facilities for small businesses and merchants,
  3. Mobile wallets and tap-and-go checkouts at supermarkets,
  4. Enabling payments through popular messaging apps like WhatsApp, and more.

Of course, traditional institutions like the bigger and older banks are less able to adapt than are fintech firms more used to innovating. For this reason, there’s been a significant increase in the number of partnerships between the two. The banks gain access to the technology they need to stay relevant to their customers while the fintech firms gain access to huge volumes of both customers and currency - everyone wins.

How can machine learning help in COVID crisis?

Online shopping is hardly a new experience, though COVID has undoubtedly increased our reliance on it since the start of 2020. However, there is a very significant difference in how we shop in a bricks-and-mortar store compared to how we shop online. Without visible displays of products, it’s a lot harder to browse what’s available, making impulse buying rarer.

Of course, there are tactics that websites can do to increase the numbers, but AI plays an important role in optimising these tactics. AI-powered chatbots are able to improve customer interactions and, from there, collect customer data that can be used to predict their individual needs and dodge their pain points.

This does not just impact the retail industry, either - the financial sector can benefit, too. Personalised products can be pitched to individual customers, including customised insurance products, alternative methods of making recurring payments as well as omni-channel payment options to help to make paying for goods and services as simple as possible.

That brings us to what is arguably the most important benefit that machine learning can bring to the payments industry in these troubled times - that of removing any and all friction for the customer in processing their payments. Even the slightest inconvenience can cause a customer to give up on a transaction, especially if they were not entirely decided on making a particular purchase in the first place.

Easing overseas transactions is an important part of this, especially with the sudden increase in online purchases. However, even domestic purchases have been accelerated and simplified using machine learning while supporting the need for social distancing and minimal physical contact. For example, facial recognition systems have helped accelerate in-person purchases while still maintaining financial security. It has also helped with electronic know-your-customer (eKYC) services, allowing secure transactions through mobile apps.

Further applications of machine learning in the financial sector include maximising process automation through chatbots and optical character recognition (OCR) systems. It can also be used to identify and manage risky situations and fraud and manage customer data on a large scale, enabling improved customer retention and more targeted marketing.

How widely has machine learning been applied?

As stated above, the COVID crisis is now over a year old and seems unlikely to be relenting in the immediate short term (at time of writing). However, it is worth noting that banks in particular tend to be quite slow to make significant operational changes. It’s hard to redirect the momentum of such large and unwieldy organisations, even with efforts to accelerate digital transformations.

The result of this has been illustrated in a survey of banks conducted by the Bank of England. They found that, while about half of the surveyed institutions expected to increase their use of machine learning and data science in response to COVID-19, only a third said that they had increased or planned to increase the number of such projects soon. Indeed, the number of applications of machine learning among the surveyed banks has remained relatively stable, increasing only slightly.

However, while banks are slow to adapt, the fintech sector is not. The number of new startups noticeably increased in 2020, with several rapidly leaping to the level of being able to compete with big banks. Such firms saw their customer bases more than double in 2020 while traditional banks were already reporting losing customers to fintech even before COVID struck. An increase in the number of partnerships between the two suggests that every effort is being made to plug the leaks, though, giving customers the best of both worlds.