Eighty-Four Per Cent Of BPO Hiring Sits In The Work AI Reaches First

Voice-based contact centre roles accounted for 5,579 of the sector's new hires in a single quarter, and they are the most exposed category there is.

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://johanosteyn.substack.com/p/eighty-four-per-cent-of-bpo-hiring?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

Concentration risk is a familiar concept in a portfolio and an unfamiliar one in a labour market. South Africa has built its largest entry point into formal employment on a single category of work, and that category is the one artificial intelligence handles best. The exposure sits in the industry’s own reporting.

CONTEXT AND BACKGROUND

The industry's own reporting sets out the exposure plainly. For the period October to December 2025, frontline voice-based contact centre agents accounted for 84 per cent of new hires in South Africa's global business services sector, or 5,579 jobs, with inbound customer services alone making up 4,053 of those positions and back office and non-voice work accounting for roughly 16 per cent. Cape Town economic growth MMC James Vos puts the sector at nearly 100,000 people in that city and about R24 billion a year to the Western Cape economy, and states openly that AI is going to take over these calls and that this is what the city does not want.

INSIGHT AND ANALYSIS

Routine inbound customer service is the application on which conversational AI became commercially adequate first. The work is high volume, structured, script-guided and text-adjacent, which is precisely the profile these systems handle well. Cape Town economic growth MMC James Vos puts the sector at nearly 100,000 people in that city and about R24 billion a year to the Western Cape economy, and says openly that AI is going to take over these calls and that this is what the city does not want. PwC’s 2026 Global AI Jobs Barometer, drawing on more than a billion job advertisements across 27 countries, finds early-career postings flatlining in highly AI-exposed sectors, while AI-exposed junior roles are seven times more likely than the least exposed to demand traditionally senior skills such as leadership and strategic thinking. A sector built on matric-level entry cannot satisfy that requirement by definition.

IMPLICATIONS

The sector understands the problem and has framed its answer around talent supply rather than exposure. BPESA chief executive Reshni Singh describes impact sourcing as the future of the industry and as future-proofing both the sector and its young talent, while Adjoa Ayivor of Shadow Global observes that the sector is growing and hiring at entry level while existing programmes are not producing enough beneficiaries able to take up the roles employers offer.

That diagnosis assumes the roles remain. Hywel George, director of investments at Old Mutual Investments, identifies the risk from the other direction, warning that the erosion of entry-level and junior roles undermines the first rung on the ladder for young people and makes social mobility harder, and that South Africa’s inequality could intensify if ownership of advanced AI systems concentrates among a small number of global firms and investors.

CLOSING TAKEAWAY

In any other setting, an exposure this concentrated would trigger a stress test. A bank with 84 per cent of new lending in one category would be required to model the downside and hold capital against it. No comparable exercise appears to have been published for the industry that is the primary route into formal employment for South African school leavers, and which has committed itself to half a million cumulative jobs by 2030. Three things follow. The association should publish an exposure assessment by service line rather than leaving the calculation to inference. Operators repositioning towards complex work should state how many seats that model supports, because higher value per seat with fewer seats is a good outcome for a firm and a poor one for a country. Incentive schemes built to reward headcount should be tested against a business model that may no longer scale with it.

Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net