27four’s Investment Management South Africa 2026 survey reports that 82% of the country’s investment managers now use AI, yet only 12% have embedded it extensively across their businesses. The gap between those two figures is the central governance question raised by the findings: how firms are overseeing the technology as their reliance on it deepens.
The survey covered 102 managers who together oversee R8.38 trillion in assets and employ 5,853 people. Of those using AI, 70% deploy it selectively in specific functions rather than across the enterprise. For regulators, boards and oversight functions, the more pressing detail lies in where the technology has been allowed to operate and where firms have held it back.
Among firms that reported their usage, 90% apply AI to investment research and idea generation, 62% to middle- and back-office operations, and 56% to due diligence. The picture changes closer to the point of decision-making. Just 21% use AI for portfolio construction or optimisation, and only 2% for trading and execution. Firms appear to have drawn their own internal boundaries, concentrating the technology in support functions while keeping humans in charge of execution.
The survey documents where AI has been accepted, not whether it has improved investment returns, and 27four is careful on that point. What the data does illuminate is how managers anticipate their workforce and control structures will adapt. Only 1% expect AI to reduce overall headcount over the next three years. Some 34% expect to reallocate responsibilities without cutting staff, 22% expect to create new roles, 28% foresee no material impact, and 15% remain unsure.
Meanwhile, when asked which functions could face automation or role reductions, respondents pointed first to operational areas. Operations and middle-office work was named by 48%, investment functions by 24%, technology and data by 23%, client servicing and distribution by 14%, and risk and compliance by 7%. A further 18% expect no material role reductions. The function charged with overseeing technology risk itself is among the least likely to be reduced, according to respondents’ own expectations.
The governance question sharpens around insurance and risk transfer. As firms depend more heavily on data, systems and third-party technology, their exposure to cyber risk grows. Disclosure patterns suggest a lag. While 89% of managers report holding crime and civil liability insurance and 81% carry directors’ and officers’ cover, only 67% disclose cyber insurance. Among firms that disclosed their limits, the median cyber cover stands at R10 million, compared with R50 million for crime and civil liability.
27four cautions that these risks are not directly comparable, but the firm argues the gap raises a fair question: whether controls, contingency plans and insurance arrangements are keeping pace with managers’ growing dependence on technology. That dependence extends to two specific oversight duties the survey highlights, namely validating analysis produced by AI tools and supervising outside technology providers on whom firms increasingly rely.
Taken together, the findings sketch an industry that has embraced AI at the research and operational edges of its business while retaining human authority over trading decisions, and that expects to reshape roles rather than shed them. The unresolved issues are ones of assurance rather than adoption: whether internal validation processes, vendor oversight and insurance cover have matured as quickly as the technology itself has spread through the sector.
The Investment Management South Africa 2026 report is published by 27four and draws on self-reported data from participating managers regarding their AI adoption, applications and expectations. The full survey is available from 27four.