This past Friday was arguably the biggest day of the year for retailers, particularly online retailers.

Throughout last week, you probably received emails about massive Black Friday discounts. Some of you might have put together wish lists to check out at the stroke of midnight while others used your phones, to scout whether a 30% discount was worth the still hefty price tags.

What you may be interested to know is that artificial intelligence (AI) has tailored your online experience. AI is a technology that makes machines smart. Black Friday, a phenomenon which began in the US after Thanksgiving, traditionally served as a means of shifting stock which had not been sold during the year, at a reduced price. But in an increasingly technological age, retailers now use your data to predict what you will spend your money on. The sneakers on your wish list may have been recommended to you by an app based on other goods you have bought before, suggestions for the latest iPhone may be based on a tweet about how listless the photos on your Android look. This makes managing pricing, inventory and distribution far more efficient for retailers.

Arguably, Black Friday is one of the biggest days for banks, and this is only possible because of their digital offering. With a single click to checkout, a card verification value (CVV) number and a one-time pin, you would probably have completed your entire purchase on a smartphone. It was estimated last year by one bank that the total number of transactions processed through its payment system over the course of the Black Friday weekend was 10 million on Friday, eight million on Saturday, 5.9 million on Sunday and six million on Cyber Monday.

It is not surprising that there is a push for a more digitised offering from banks. The banking sector, arguably, has seen the most significant overhaul with the Fourth Industrial Revolution (4IR). This, of course, has been coupled with immense fear. As branches become somewhat obsolete, there are increasing worries around job losses. Threats of a strike earlier this year all but confirmed this, after several retail banks announced retrenchments. I, for one, have not visited a branch of my bank this year because I can perform everything I used to do at my physical branch on my phone app. I also realise that this phone banking application is not perfect, as it is only available 80% of the time due to technological challenges.

But to stay ahead of the game and meet customers’ needs, banks cannot afford to pay for costly and largely underused branches. Instead, the focus needs to shift to improving their online offerings. According to the Business Insider Intelligence Report on AI in Banking, “The aggregate potential cost savings for banks from AI applications is estimated at $447-billion by 2023, with the front and middle office accounting for $416-billion of that total.”

One of the central functions of banks is to use internal data such as transactions, purchase history and patterns, as well as social media and mobile banking usage to provide better customer service, detect fraud and understand consumer sentiment, for example.

The reality of the 4IR is that the nature of the workforce has to change to adapt to the era. There are three changes in jobs that will come as a result of the 4IR. First, some jobs will disappear altogether. Second, some jobs will change. An example of this is the medical profession, where health workers will progressively require proficiency in technology. Third, new jobs will emerge. For instance, banks now hire people with new job titles, such as the chief artificial intelligence officer, which did not exist a few years ago.

Despite automation at the branches and an increasing emphasis on using banking apps, South Africa’s six largest banks had 152,441 employees in 2018 – an increase of almost 4,000 from 148,500 in 2015, according to the Banking Association SA (Basa). The caveat, of course, is that there is a demand for different skills. The bankers of the future must have some understanding of technology, society, finance and accounting.

In a thought piece for Quartz earlier this month, Antony Jenkins, the founder and executive chair of 10x Future Technologies said, “The bank of the future will need to think about people in a completely different way. This will require a customer-first approach that is deeply embedded in the process of how services are designed and delivered. It’s an approach that is digitally driven but focused on solving our human problems and needs.”

In South Africa, for example, we have to tackle the issue of financial inclusion. FinMark Trust in 2018 found that 90% of people in South Africa have access to financial services, and 80% have bank accounts. Nevertheless, many of these bank accounts are not active, and the majority of South Africans use informal financial services, many of which are unregistered.

According to journalist Tehillah Niselow, more than 275,000 people now use uKheshe, a micropayment platform that was established in 2018. This platform, which was recently piloted to car guards in the trendy Joburg suburb of Linden, allows you to accept an uKheshe payment by buying a card for R20 at a money counter at Pick n Pay, where your ID can be verified on the spot. You do not require a smartphone to accept money, just a mobile phone that is USSD enabled. The subscription cost of this service is R5 per month plus R5 for a withdrawal from any Pick n Pay counter.

Elsewhere, small businesses are making use of mobile payment platforms that require a more scaled-up cellphone such as SnapScan, Zapper or a mobile banking app-enabled with a QR code. This is significant when you consider that smartphone penetration in South Africa in 2018 was nearly double that of 2016 at 81.72%, according to the Independent Communications Authority of South Africa. A study by Nielsen suggests that South Africa is one of the countries with the highest self-reported rates of participation in mobile banking, in the company of China, India and Sweden.

Despite the shift towards faster and more inclusive banking, there are also ethical considerations to take into account. The risk of developing biased data sets which can significantly affect application processes or recommendations is of particular worry in a country as unequal as ours. If we consider that AI mimics human intelligence, it is a given that in some instances it picks up on human bias.

Take, for example, Amazon’s AI recruiting tool which had to be scrapped after it showed bias against women. The device reviewed job applicants’ curriculum vitae (CVs) to automate searching for top talent. However, in 2015, the company observed that the system was biased because of the dominance of men in the tech industry. The device learned to prefer male over female candidates by penalising CVs that included words such as “woman,” and “female soccer team”. Here, it is essential to ensure that banks have a team that is diverse in skills, backgrounds, and experiences to ensure that models are exposed to a diverse set of data. DM