AI Is Not Automation: Why Behaviour Change, Not Hype, Will Decide The Future Of Work

South African businesses are racing to deploy artificial intelligence (AI), but many risk confusing automation with true AI with potentially costly consequences. Speaking at the Knowledge Resources Conference in Johannesburg on 20 August, Geoff Cohen, Partner at digital transformation consultancy, DY|DX, urged delegates to take a step back and interrogate whether they are solving the right problems.

Cohen warned that ‘automation is not AI, and AI is not automation’, and stressed that without the right organisational behaviours and processes in place, technology adoption often fails to deliver the promised value.

“Automation is a process, while AI is a tool,” Cohen explained. “One without the other rarely achieves lasting results. The real transformation happens when organisations design new processes that take advantage of AI’s potential and that requires behaviour change across teams and leaders. At DY|DX, our strength lies in combining technology with change management to help businesses build these new processes effectively.

“Too often, businesses are sold a vision of tools that promise efficiency and cost-cutting disguised as AI,” says Cohen. “But simply automating existing processes does not make an organisation more intelligent. True AI adoption requires leaders to rethink how work gets done, how people make decisions, and how behaviours must shift to support new systems. Otherwise, you end up with faster processes and more activity, but not better outcomes.”

Cohen argued that healthy skepticism is now a critical leadership skill in the age of AI hype. Far from being negative, skepticism protects businesses from rushing into poor decisions. He advises executives to interrogate every AI proposal with three deceptively simple questions:

1. What problem are we actually solving?
2. What evidence do we have that this will work?
3. How will we know if we’re wrong?
   

“South Africa’s corporate landscape is noisy,” Cohen added “There’s no shortage of vendors promising transformative AI. But skepticism is the cognitive bodyguard leaders need to separate the signal from the noise.”

To help organisations avoid blind adoption, DY|DX has suggested the ‘Three Zones of AI’ framework, which Cohen presented to delegates:

• Zone 1: Offload – Automate predictable, rules-based tasks. But critically, ensure displaced human capacity is redirected into higher-value roles through reskilling.
• Zone 2: Amplify – Combine human judgment with AI insights for more complex decision-making. This requires building enterprise-wide AI literacy and clear governance protocols to ensure accountability.
• Zone 3: Explore – Use AI to create entirely new ways of working, from innovation pilots to emerging roles and products that wouldn’t exist without human-AI collaboration.
  

“The danger is that businesses stop at Zone 1 where they chase efficiency and cost-cutting,” Cohen explained. “This creates what we call ‘trapped costs’. if you make a task 30% faster through automation or AI, what are you actually doing with that productivity gain? Unless managers intentionally redirect that capacity into higher-value work, the benefit is lost. The real opportunity lies in Zones 2 and 3 – amplifying human decision-making and exploring new possibilities that fundamentally change the way organisations create value.”

Cohen emphasised that the success of AI adoption is less about technology and more about how people behave within new systems.

“Process change hinges on behaviour change,” he said. “If people don’t trust the outputs, don’t know how to ask the right questions, or don’t adapt their ways of working, then even the best automation or AI project will fail. Technology can only take you so far, the rest is human capability, culture, and governance.”

He called on HR leaders and change managers to take a central role in AI adoption, ensuring that employees are equipped with not only technical literacy but also the critical thinking skills to work alongside AI responsibly.

Cohen concluded that businesses need to broaden their success metrics when it comes to AI as efficiency gains are only part of the story. 

“Organisations must ask: Are our people more engaged? More creative? Are we building resilience, adaptability, and innovation capacity? If the answer is no, then AI hasn’t transformed your business, it has simply automated it.”
To find out more about DY|DX and the services offered, head to their website.