Nearly two-thirds of South African consumers have used artificial intelligence to assist with shopping-related tasks, according to new survey data from Visa pointing to a rapid shift in how South Africans interact with digital payment technologies.
The Visa Spending Shift Survey, conducted by Morning Consult among 1,000 South African adults in October 2025, found that 63% of respondents have used AI tools for tasks including product research, brainstorming gift ideas and communicating with customer service. Almost 45% believe AI-powered tools are more likely to find the best possible price.
Despite growing adoption, a clear trust gap remains: 63% of respondents still prefer speaking with a human customer service representative over an AI chatbot, and 60% expressed concern about how their personal data is being used by AI-powered shopping and payment tools.
The survey also highlights growing comfort with biometric authentication, with 66% of respondents saying they already use fingerprint or facial recognition to authorize payments. Digital wallets are gaining ground as well, with 39% of South Africans identifying them as the fastest payment option.
Among the most striking findings for the cross-border payments market, 57% of respondents who have sent money abroad said they would consider using stablecoins as a form of payment, suggesting an appetite for blockchain-based alternatives in the remittance sector.
Security remains the dominant concern overall, with 88% of respondents rating it as extremely important when choosing a payment method. Consumers view banks (76%) and payment networks (71%) as the primary line of defense against emerging threats, though 67% believe they bear significant personal responsibility for detecting fraud.
“South Africans are showing that trust and technology can go hand in hand, with consumers increasingly willing to adopt new payment solutions when they deliver both convenience and peace of mind,” said Lineshree Moodley, Visa’s South Africa country manager.
The survey was conducted online with data weighted by gender, education, age and race. Results carry a margin of error of plus or minus 3 percentage points.
