Kathy Gibson reports from the IITPSA Workforce Conference – The fear that AI will lead to wholesale job losses is front and centre of the current debate.
But as AI comes of age, and organisations start implementing it into their workforces, we have an opportunity to reframe the debate, says Dr Pragashani Reddy, chief commercial officer of Digital Solutions Group (DSG).
“AI lets us redesign the work better,” she points out.
A key consideration when implementing AI is that employers can’t treat jobs simply as a job title. The reality is that a job role is a bundle of tasks – and AI changes the bundle before it changes the job title.
Reddy believes that the real disruption is the task, and what we do with the time it returns.
It’s important to remember that automating a task doesn’t automatically improve the process. “We need to determine deliberately what needs to be automated – and it is definitely not everything.”
Results from the IITPSA Skills Survey show that 80% of respondents report a lower workload and or task complexity from implementing AI. The question is whether these gains translate into business outcomes, Reddy says.
The jury is out on the impact that AI will have on jobs. More than half of respondents already use AI in their daily use and report productivity gains, but 47,9% are both optimistic and concerned about work implications. Only 6% believe they face a high AI-related job loss risk.
The paradox here is that AI can reduce the work, but simultaneously raise the bar, Reddy says. This is because the opportunity is not necessarily automation, but is rather dependent on how the organisation redesigns work.
AI can help to shrink routine tasks, help with first drafts, develop manuals, do status reporting and simple ticket triage.
But what is more valuable is verification and judgement, system design and integration, customer domain understanding, security, governance and accountability.
“We need to redirect the productivity we gain to innovation, architecting and resilience,” Reddy says. “Because without a deliberate redesign, AI could just mean higher workloads.”
As organisations introduce AI into the workplace, it is vital that humans retain control. For GenAI, the involves directing each next step; for agentic. AI or workflow agents are good at using approved tools, checking progress and handling exceptions, but it needs humans to be in charge of permissions, monitoring and escalation.
Reddy outlines how six different ICT job categories could be changed by the adoption of AI.
- Software development: AI could be used for boilerplate code, with humans responsible for architecture, security design and code review.
- Service operations: AI would implement tickets and triage, but humans need to handle complex incidents, knowledge and design.
- Cybersecurity: AI is good at alert summarisation, with humans taking on threat defence and adversarial testing.
- Data and analytics: AI could do dashboard production, with humans responsible for data quality, internal relations and decision-making.
- Product/project management: AI does the status reporting, but humans do the orchestration, adoption and outcomes.
- Cloud/infra: AI can handle routine provisioning, humans provide the resilience, AI operations and cost controls
In light of the reality of the skills shortage, South African organisations have the potential to redesign tasks.
The current talent bottlenecks are experience (24%), education and training (29%), lack of diversity (16%) and the brain drain (12%).
“The biggest gap is not just technical – it is experiential,” says Reddy. “The critical risk is that we automate entry-level tasks without replacing the learning pathway. This risks weakening future talent pipelines.”
Competing for talent
What is ironic in the midst of South Africa’s unemployment crisis is that organisations are still importing talent.
The IITPS Skills Survey points to a reliance on overseas capacity alongside persistent local shortages, with 52% of organisations outsourcing IT functions overseas and 28% looking overseas for specialised skills.
“The strategic concern is whether we are developing local expertise,” says Reddy. “I employers can create opportunities for local resources, there would be no need to constantly look for international expertise.
“We need to look to meaningful participation.”
Emerging careers
As AI permeates the organisation, it’s important to realise that not every employee need to be an AI researcher.
“Every sector needs people can apply AI responsibly,” Reddy says.
New roles include AI workflow/agent designer who maps tasks, tools and guardrails; the AI product leader who connected user need to value; an AI assurance specialist who tests bias, safety and reliability; a data/knowledge steward who makes information usable; an AI security specialist who ensures that models and systems are protected; and a human-AI experience designer who would design interface and collaboration.
From hype to execution
Reddy recommends the organisations adopt a 90-day workforce reset that maps the work, proves the value, and redesigns the roles.
The first 30 days should be dedicated to listening and mapping. This mapping should be of tasks – not job titles. This time would also be used to set risk boundaries, and baseline quality and time.
Day 31 to 60 would be to pilot and learn. Reddy suggests tunning two to three low-risk pilots, train people to verify, measure and reword tasks, and build trust.
The final 30 days would be dedicated to scale and redesign. “Scale only what works,” Reddy says. During this time, should also fund reskilling and redesign people’s career paths.
“The future is not pre-written,” she concludes. “AI will change work, but we can shape what the work becomes.
“The goal is not simply a workforce with fewer humans; it is a workforce with more capability, opportunity and better tools.”