No longer the stuff of science-fiction, Artificial Intelligence (AI) has captured the attention of South Africans across the board. In fact, it would be difficult to scroll any news or social media platform without finding at least one reference to the likes of ChatGPT or DALL-E.
These platforms have wowed users for their ability to synthesise large amounts of data and generate written articles and digital artworks in seconds. More importantly, they have sparked widespread discussion (and fear) around the potential impact that AI will have on a range of industries – from those in creative fields, to finance, engineering, and more.
The payments industry has not escaped this debate. While opponents worry that automation and machine learning will lead to job losses, proponents believe that the technology will act as a great enabler, ushering in the next wave of improvements to the way that consumers and businesses alike make transactions and manage their financial lives.
The industry has already begun to see the latter emerge, with innovations such as AI chatbots enabling a more proactive and deeply personal form of customer experience.
Beyond this, AI also promises to improve efficiencies , remove errors and lower costs by automating routine processes, as well as unlock new and previously unrealised opportunities based on an improved ability to process and generate insights from vast amounts of user data, gathered via digital payment channels.
Evidently, AI can go a long way towards addressing several challenges in the payments ecosystem – and more so in South Africa, where the industry continues to grapple with limited access to traditional banking services, low levels of financial literacy, and a lack of trust in digital platforms.
However, it would be naïve to believe that the technology will revolutionise the industry overnight. Rather, there are several challenges which first need to be overcome for AI to play a larger role in the South African payments landscape.
Cash will always be king (for now)
A primary challenge facing AI in South Africa will be the region’s reliance on cash, which will continue to be a primary form of exchange as it allows immediate participation in the economy for all, including people living in rural areas or foreign nationals without formal documentation or bank accounts.
In this regard, cash is unmatched in its ability to provide speed, flexibility, anonymity, and affordability as legal tender. It is also this physical nature of cash that inhibits the uptake of AI, which relies on large amounts of data, which in turn can only be generated at mass through digital transactions.
When consumers make payments online, the businesses that facilitate and process these transactions are able to track a myriad of data points – from the date, time, location and amount paid, to more granular details such as the amount of time a user spends at a specific step of the payment process.
Together, this information can be processed by computer algorithms to paint a very detailed picture of the user’s experience and inform the payment provider on how to adjust their systems.
High data costs will lead to less data generation
Digital payments have, however, seen a surge in popularity in South Africa, thanks in part to the uptake of fintech in recent years, the proliferation of mobile devices, and the COVID-19 pandemic, during which consumers became more accustomed to transacting online.
In fact, data from the World Bank has shown that, during the pandemic, more than 40% of adults in low and middle-income economies like South Africa made a payment online for the first time. In doing so, South African consumers have greatly expanded the potential for data generation – but this is unsustainable if high data costs persist.
As the price per megabit of data has decreased in developed countries, a combination of lower income levels, limited infrastructure, and a lack of competition in the telecommunications market – coupled with high taxes and tariffs on mobile devices and services – has seen data prices remain relatively high in Africa.
This is no different in South Africa, where consumers pay an average R78.50 per gigabyte data while their counterparts in the United Kingdom pay only R37.12 for the same amount. This will in-turn force many consumers to resort to more affordable payment channels, such as cash.
In an effort to overcome high data costs, many retailers and merchants have looked to alternatives such as zero-rating their marketplaces or investing in Unstructured Supplementary Service Data (USSD) services to make it easier for consumers to transact online.
Good examples can be found in organisations such as Pay@, a leading payments aggregator in South Africa, offering merchants and consumers alike a range of payment methods via integrations in the zero-rated Capitec Banking App and MTN Momo, among others.
By making it more affordable for users to transact online, these platforms are able to generate data which can be used to power AI.
SA’s energy crisis isn’t helping
Regardless of data costs and consumer habits, the uptake of AI in the South African payments industry cannot escape the current energy crisis. While mobile phones rely on batteries, sustained and high levels of loadshedding will affect their rechargeability and longevity, with outages also impacting the speed, coverage and performance of telecommunications infrastructure.
Together, this severely limits the use of digital devices and the ability of consumers in a mobile-dominated market to make payments online.
Unable to use their electronic devices, consumers will have no other option than to fall back to traditional means of exchange – such as using cash in physical retailers – who are able to maintain payment systems by providing generators and other forms of electricity supply.
Together, the interlocked challenges associated with long-standing consumer preferences, high data costs, and an ongoing energy crisis present significant barriers to the short-term uptake of AI in South Africa.
But this is not to say that the technology will not break new ground in 2023. Rather, players in the payments sector can expect AI to take subtle steps towards optimizing internal processes and workflow, before being implemented on a customer-wide basis.
By Carel Botha, Business Development at Pay@