BANKING ON THE ARTIFICIAL INTELLIGENCE REVOLUTION

Publication Date: 28/08/2018

On April 21 1981, Standard Bank introduced 25 ATMs around Johannesburg, one of the first rollouts of ATMs by a major bank.
At first the IBM 3624 ATMs operated only from 7am to 7pm, Monday to Saturday, and it took 20 seconds for a cash withdrawal, while you would have to wait 60 seconds for a balance inquiry.
Thirty-seven years later there are considerably more ATMs around the country, but you can also do your banking at any time on any day on just about any device — while online or on an app on your phone or even while on social media, all within seconds.
Those ATMs could be considered an early form of fintech; a bank spokesperson told the Rand Daily Mail at the time that they would ultimately perform 90% of teller duties.
The introduction of ATMs were a disruptor, shifting customers from the bank to outside where they engaged with technology to get their money that before would have been accessed from inside the branch.
The main thrust is that banks are looking into AI as a way to cut costs
Recently, a similar shift has taken hold as South African banks have embarked on digitalisation.According to a Citi GPS report on the Bank of the Future released this year, globally banks are the biggest investors outside the tech space in artificial intelligence (AI). The banking securities sector poured $1.9bn into AI in 2016 and this is expected to jump to $7.5bn (R107.5bn) in 2019.
Co-Pierre Georg, an associate professor at the African Institute of Financial Markets and Risk Management at the University of Cape Town, said that he expected to see two major developments in the coming years.
The first is the emergence of online banks like BankZero and TymeDigital as branches increasingly become a relic of a time before smartphones. The second is disintermediation as more will be possible without the help of intermediaries, starting from payment services, forex transactions and even parts of the loan book.
Georg said AI had already changed banking and would continue to do so.
“There are some applications, such as robo-advisers, where AI improves the customer experience. But the main thrust is that banks are looking into AI as a way to cut costs.
“The effect will be that a large number of jobs in the financial services industry will be lost.”
The current system was inefficient and the cost of compliance enormous. Automation and big data were expected to solve some of these inefficiencies, but the consequence was that fewer people would be needed for the same services.
During Standard Bank’s results presentation, CEO Sim Tshabalala said the cost of a transaction on digital channels was much less for clients than in a branch or at an ATM, but that was putting pressure on revenue.
“The volumes are increasing but the pricing is much less on electronic channels … [yet] the cost of product and service delivery is still the same because we haven’t fundamentally changed the infrastructure.”
There will be a big drive towards customer self-service on the latest social media platforms Arrie Rautenbach, CEO of Absa’s Retail & Business Banking.
But Mark Brits, a senior GM at the Banking Association SA, said while there would be job losses there would also be new jobs created.
Brits said it was difficult to make a prediction on job losses, as a large portion of the country still favoured cash and many South Africans did not have smartphones and were therefore excluded from the new system.
Raj Makanjee, CEO of FNB Retail, said the bank’s aim was to complement and enhance rather than replace human capital.
“We invest in our people to take advantage of technology and automation for a better service.”
Over the years, IT-based roles at FNB have increased significantly, with IT and IT-related roles accounting for close to 15% of the workforce.
Makanjee said automation offered significant improvement in areas where repetitive tasks could be performed optimally by machines that learn and self-improve.
He said that, for instance, in areas such as judgmental credit (when an assessor approves or declines a credit application), the introduction of automated credit decision models resulted in the less complex decisions being automated and that freed the judgmental credit assessors to focus on the more complex matters.
Brits said in the next five to 10 years he expected SA to still have a mixed offering of digital and brick-and-mortar branches, but there would be a reduction in branch networks and more ATMs.
But branch network reduction has not yet taken hold as a number of banks increase the number of locations.
Capitec last year bought a 40% stake in Creamfinance for à21m (R348m) that will form part of the group’s online lending capabilities and offer advanced credit scoring through machine-learning capabilities. The group is also rolling out 15 branches without cashiers.
Capitec said it was not replacing employees, but had consultants to assist clients with more complex transactions.
The group said its view on automation and AI was that it was not meant to replace people, but rather enhance human capabilities.
Absa, which recently launched ChatBanking on WhatsApp and Samsung Pay, said its technology unit had invested about R350m in the past three and a half years in automation technologies, specifically robotics and AI, which include voice and face recognition as well as fraud detection.
Arrie Rautenbach, CEO of Absa’s Retail & Business Banking, said, a big portion of Absa’s workforce would be supplemented by AI and other supplementary automation technologies in the future.
“There will be a big drive towards customer self-service on the latest social media platforms and bolstering existing channels, enabled through integration with voice assistants such as Alexa or Siri and augmented and fully virtualised banking services,” said Rautenbach.