The Agentic AI Era Demands New Entry-Level Roles. Are SA Companies Creating Them?
With all the speculation about AI eliminating entry-level jobs, it is no wonder young South Africans are anxious about launching their careers.
This anxiety is entirely understandable. According to the latest Quarterly Labour Force Survey from Statistics South Africa, the official unemployment rate stood at 31.4% in the fourth quarter of 2025, with 7.8 million people unemployed. Even graduates are not insulated, with the graduate unemployment rate at 10.3%, and year-on-year comparisons show that rate is rising, not falling. Into this already strained labour market, AI is now automating the very entry-level tasks that have historically been the first rung on the ladder.
This moment demands honesty from business leaders: yes, AI is eliminating some traditional entry-level tasks, and yes, the path into professional work is changing. But as with the dawn of the internet before it, AI does not have to mean fewer opportunities, if companies commit to creating new roles that pair human judgement with AI capability. The shift is not from employment to unemployment. It is from execution to judgement. And South Africa, with its youthful population and its urgent need for meaningful job creation, has more to gain from getting this right than almost any economy on the continent.
Mastering AI partnership, from tool to catalyst
AI fluency is not about understanding algorithms, it’s about mastering a new kind of collaboration that transforms how work gets done, progressing from using AI as a basic tool for simple tasks, to handing off entire workflows, engaging it as a thinking partner that challenges assumptions, and ultimately to using it as a catalyst that surfaces strategic opportunities you had not yet considered.
This progression is already reshaping what employers need in South Africa specifically. According to PwC South Africa’s analysis of the 2025 Global AI Job Barometer, occupations most exposed to AI in South Africa have seen skills requirements change at 1.32 times the rate of roles less affected by the technology. Demand for AI skills in the education sector surged from 4.9% of job postings in 2021 to 8.5% in 2024, while the ICT sector saw AI skill requirements rise from 5.5% to 7.9% over the same period. Critically, augmentation-exposed roles, those where humans and AI work together, have shown an average 20% growth rate in job postings across sectors, compared to a 2% decline in postings for pure automation roles.
The challenge for South Africa is that this global pressure lands on a labour market already under severe strain. Building AI fluency into the professional experience from day one is not something the education system can do alone. It is something business leaders must own.
AI. Image source – Canva generated
Elevating human judgement as the competitive edge
Most organisations are still hiring for yesterday’s workforce, seeking people who can execute tasks efficiently. In a world where AI executes tasks near-instantly, judgement separates high performers from the rest.
The emerging skill set centres on evaluation and direction: assessing whether an AI-generated client proposal reflects the right brand tone and ethical positioning, spotting the gaps in an AI-produced financial model, and asking the follow-up questions that unlock new strategic directions. These are not technical skills. They are human skills, elevated by AI collaboration.
The World Economic Forum’s Future of Jobs Report 2025 projects that 170 million new roles will be created globally this decade, and the roles growing fastest are precisely those that require humans to direct, evaluate, and exercise ethical oversight over AI systems.
South Africa’s context makes this more urgent, not less. If the judgement-based, AI-augmented roles that define the agentic era are designed only for those who already have access to quality education and digital infrastructure, the technology simply reinforces South Africa’s existing geographic and socioeconomic divides. That is a choice, and one business leaders make by design or by default.
Redesigning the professional rite of passage
When AI handles data gathering, report formatting, scheduling, and first-draft creation, the rationale for confining new hires to those tasks disappears. And when AI handles execution, new hires engage with strategic thinking from day one. Success now centres on impact and contribution, not on proving yourself through months of routine tasks.
For this to happen, leaders must build formal AI fluency programmes that measure genuine capability, not just certification completion, and create clear pathways from South Africa’s universities and TVET colleges into these new roles.
The private sector has already demonstrated it can move decisively when it commits to doing so. The Youth Employment Service created over 209,000 quality work experiences since inception, contributing R12.3 billion to the economy through youth salaries alone, powered entirely by private sector participation with no state funding. The same energy now needs to be directed at redesigning what those roles look like inside the agentic enterprise.
Accountability in the Agentic era
For business leaders, this transformation requires more than adaptation. It demands accountability. We must actively redesign entry-level roles to emphasise judgement and orchestration over task completion. We must create clear pathways into these roles, not leave young professionals to figure it out alone. Most importantly, companies must commit to creating these positions, not just talking about creating them.
The entry-level job of the future is not about knowing the most. It is about thinking the best. It is about being the person who asks “Should we?” not just “Can we?” South Africa has no shortage of people capable of that thinking. What it needs are business leaders committed to building the roles that let them prove it.
- Ursula Fear, Senior Talent Program Manager at Salesforce
END
You May Also Like
Lovable AI’s Astonishing Rise: Anton Osika Reveals Startup Secrets at Bitcoin World Disrupt 2025
The Nationwide Tug-of-War over Prediction Markets
