South Africa’s banks are taking markedly different paths when it comes to artificial intelligence, and the divide says as much about their client bases as it does about technology. Investec’s chief information officer, Graeme Lockley, made it clear that the specialist private bank is resisting the AI‑first rush seen at rivals such as Discovery and Capitec, insisting that its edge lies in people rather than algorithms. Speaking from the bank’s Sandton headquarters during a careers‑in‑tech event for Grade 10‑12 learners, Lockley framed the day as a live demonstration of Investec’s belief that relationships, not machines, define the institution. He argued that while AI can handle repetitive tasks, it cannot replicate the nuanced judgment that comes from long‑term client connections.
AI strategy diverges among South African banks
Lockley warned that treating AI as an infallible decision‑maker is a recipe for error, likening blind trust in the technology to the early days of search engines when users were wary of Google’s results. “If you think of AI as the super smart savant, you are going to make some terrible choices,” he said, adding that the tool is most useful when it supports, rather than replaces, human critical faculties.
Discovery Bank, by contrast, has placed AI at the heart of its growth plan, launching Vitality AI earlier this year. The division posted an operating loss of R155‑million for the six months to end‑December 2025, more than double the prior period, reflecting the scale of the investment. The bank draws behavioural and health data from over 40 million Vitality members across more than 40 markets, which it claims gives it a genuine advantage in deploying AI‑driven insights.
Capitec has gone a step further, putting AI tools into the hands of almost 5 000 employees and running an agentic AI system inside its business‑banking credit processing operation. Chairman Santie Botha and CEO Graham Lee told shareholders in a joint letter in April that “AI is not a future aspiration. It is already at work.” The bank’s 2026 integrated annual report shows that AI‑driven models blocked more than 131 000 fraudulent beneficiaries, prevented over 394 000 scam payments and averted roughly R673‑million in potential client losses over the year. Lee emphasised that the aim is not to cut staff but to limit headcount growth while serving a expanding client base.
Even the country’s purely digital challengers are echoing a similar sentiment. GoTyme CEO Cheslyn Jacobs noted that customer behaviour has shifted, with clients no longer loyal to a single institution but picking products to suit specific needs and often maintaining multiple banking relationships. GoTyme is positioning itself to capture that multi‑banked trend at scale, using AI to enhance service offerings rather than simply to automate processes.
Across the sector, none of the major banks is presenting AI as a route to a smaller workforce. Instead, the technology is being sold as a means to boost efficiency, improve risk management and free staff for higher‑value, human‑centred activities. As we continue to monitor these developments, it is clear that the conversation in South Africa’s financial services is less about man versus machine and more about how each institution chooses to blend the two to best serve its customers.