Why South African workers must become supervisors of digital labour
Loading player...
ADG’s Cliff de Wit on why the next management challenge is overseeing teams of autonomous software agents.
Not long ago, organisations were experimenting with generative AI to draft e-mails, summarise documents and answer questions. Today agentic AI is changing what the technology is for.
In a conversation on TechCentral’s TCS+ podcast, Cliff de Wit, MD for South Africa and group chief innovation officer at Accelera Digital Group (ADG), argued that the real story is not machines replacing people but a fundamental shift in how work gets done — and where humans create value.
Rather than completing individual tasks on command, AI agents can be given an objective and left to determine how best to achieve it. De Wit illustrated this with a know-your-customer (KYC) use case in which agentic AI is already automating complex workflows and accelerating business outcomes.
AI is no longer simply a tool that assists workers. Increasingly, it is a digital workforce in its own right.
In a country with one of the world's highest youth unemployment rates, concerns about automation are understandable. But De Wit challenges the idea that AI should be viewed purely as a job-replacement technology.
"The human skills involved in business are fundamentally changing, but the basics of business still remain," he said.
Businesses will still need people to solve problems, exercise judgment, build relationships and drive innovation. What changes is how those outcomes are achieved: as routine and administrative work is automated, workers move up to activities where uniquely human skills matter most.
The sharper risk is distributional. Without investment in education, training and digital skills, AI's productivity gains could widen existing inequalities — not because jobs disappear, but because opportunity concentrates among those who know how to work with the technology. AI literacy may soon matter to employability as much as computer literacy did during the digital revolution.
Organisations are already deploying agents into operational environments, in some cases slotting them into workforce structures alongside human employees. Leaders are no longer managing only people; they are managing mixed teams.
De Wit believes businesses are better prepared for this than they think. Companies have long used governance frameworks and risk controls to oversee human decision-making, and many of the same principles adapt to digital workers. Whether a decision is executed by a person or a system, accountability, escalation and oversight remain essential.
Which decisions should be delegated, then, and which should stay under human control? De Wit's answer comes down to risk, and he sets out a framework for categorising it. The future, on his reading, is unlikely to be fully autonomous. Successful organisations will run human-in-the-loop models in which AI handles execution while people supply direction, judgment and accountability.
De Wit is optimistic. Many local organisations are still experimenting, but a growing number are moving past pilots to solve real problems and generate measurable value.
For leaders, the question is no longer whether to adopt AI, but how to do so responsibly, effectively and inclusively. The future workforce will consist neither solely of humans nor solely of machines — it will consist of humans who know how to lead, govern and orchestrate digital labour.
And that future is arriving faster than many realise. Don’t miss a great discussion!
Not long ago, organisations were experimenting with generative AI to draft e-mails, summarise documents and answer questions. Today agentic AI is changing what the technology is for.
In a conversation on TechCentral’s TCS+ podcast, Cliff de Wit, MD for South Africa and group chief innovation officer at Accelera Digital Group (ADG), argued that the real story is not machines replacing people but a fundamental shift in how work gets done — and where humans create value.
Rather than completing individual tasks on command, AI agents can be given an objective and left to determine how best to achieve it. De Wit illustrated this with a know-your-customer (KYC) use case in which agentic AI is already automating complex workflows and accelerating business outcomes.
AI is no longer simply a tool that assists workers. Increasingly, it is a digital workforce in its own right.
In a country with one of the world's highest youth unemployment rates, concerns about automation are understandable. But De Wit challenges the idea that AI should be viewed purely as a job-replacement technology.
"The human skills involved in business are fundamentally changing, but the basics of business still remain," he said.
Businesses will still need people to solve problems, exercise judgment, build relationships and drive innovation. What changes is how those outcomes are achieved: as routine and administrative work is automated, workers move up to activities where uniquely human skills matter most.
The sharper risk is distributional. Without investment in education, training and digital skills, AI's productivity gains could widen existing inequalities — not because jobs disappear, but because opportunity concentrates among those who know how to work with the technology. AI literacy may soon matter to employability as much as computer literacy did during the digital revolution.
Organisations are already deploying agents into operational environments, in some cases slotting them into workforce structures alongside human employees. Leaders are no longer managing only people; they are managing mixed teams.
De Wit believes businesses are better prepared for this than they think. Companies have long used governance frameworks and risk controls to oversee human decision-making, and many of the same principles adapt to digital workers. Whether a decision is executed by a person or a system, accountability, escalation and oversight remain essential.
Which decisions should be delegated, then, and which should stay under human control? De Wit's answer comes down to risk, and he sets out a framework for categorising it. The future, on his reading, is unlikely to be fully autonomous. Successful organisations will run human-in-the-loop models in which AI handles execution while people supply direction, judgment and accountability.
De Wit is optimistic. Many local organisations are still experimenting, but a growing number are moving past pilots to solve real problems and generate measurable value.
For leaders, the question is no longer whether to adopt AI, but how to do so responsibly, effectively and inclusively. The future workforce will consist neither solely of humans nor solely of machines — it will consist of humans who know how to lead, govern and orchestrate digital labour.
And that future is arriving faster than many realise. Don’t miss a great discussion!
