By Alberto Lemma
The IMF’s October 2026 World Economic Outlook highlights how AI-related investment is supporting global growth, while the World Trade Organization (WTO) recently upgraded its 2026 trade growth forecast from 1.9% to 3.9%, citing AI-driven demand, despite the ongoing Middle East conflict. However, whether AI adoption can deliver comparable gains in productivity and quality employment in African economies remains much less certain.
AI is already reaching African businesses. A World Bank survey conducted in early 2026 estimates that 42% of formal firms in Kenya use some form of AI – close to the 43% recorded in the US. These figures cover registered establishments with at least five workers, rather than the wider informal economy. Nevertheless, they raise a fundamental question: will more productive businesses generate better employment opportunities?
Kenya’s labour market underlines this dilemma. Recorded employment, excluding small-scale agriculture, increased by 822,100 between 2024 and 2025; crucially, 87% of that increase occurred in the informal sector.Previous ODI Global analysis found that approximately 15% of Kenyan workers had moderate exposure to generative AI. Recent World Bank estimates suggest that while 16.2% of jobs in lower- and middle-income countries have some exposure to AI, only 4.5% face outright automation. However, limited exposure offers little reassurance if workers remain confined to poorly paid informal jobs while opportunities to enter higher-productivity formal employment become scarcer.
This is a structural transformation problem. Productivity rises when workers produce more within existing resources, but also when labour shifts into higher-productivity sectors and occupations. AI could accelerate the first while constraining the second. In Kenya’s business process outsourcing (BPO) industry, customer support and digital services have long provided formal entry-level jobs for young workers. As AI automates routine enquiries, firms can serve more clients without additional recruitment. Tanzania’s pre-AI growth trajectory illustrates this distinction: between 2014 and 2021, structural reallocation acted as a drag on growth (-5%), meaning total productivity gains were driven entirely within existing sectors. AI risks institutionalising this pattern by dampening cross-sectoral job creation.
Recent data shows these adjustments unfolding in real time (Figure 1). Across six developing economies surveyed by the World Bank, 20.8% of AI-adopting firms reported hiring additional workers, while 7.4% reported layoffs and 12.8% froze or slowed recruitment. Falling call-centre employment across 12 middle-income countries –including a 6% decline in Kenya and 10% in the Philippines – has been driven primarily by reduced hiring rather than dismissals. Furthermore, only 10.8% of workers departing these roles transitioned into higher-paying technical or data positions.
Figure 1: AI adopters report new hiring as well as layoffs