African enterprises score higher on PwC’s new AI Fitness Index than their European and North American counterparts — yet still trail the global AI leaders on every meaningful measure of return, raising questions about whether the continent’s largest companies are scaling AI fast enough.
PwC’s newly published AI performance study surveyed 1,217 large companies globally, 85 of them in Africa. “This gap suggests that the challenge is not adoption, but execution at scale,” the firm said.
On the headline score, Africa lands at 5.7 out of 10 — just behind Asia and the Middle East, both at 5.8, and ahead of Europe at 5.5, North America at 5.1 and Latin America at 4.8. The ranking is unusual for a region routinely described as a digital laggard, though the African sample is not broken down by country and the figures will inevitably reflect the continent’s largest and most digitized economies more than the rest.
African firms are also running AI pilots at roughly the same rate as the global frontrunners — 82% versus 88% for AI leaders. The gap, in other words, is not in starting projects but in what happens after the pilot. The most AI-fit companies generate 7.2 times greater AI-driven performance than the rest on an industry-adjusted basis, and the top 20% capture 74% of all AI-driven financial returns. Africa sits on the wrong side of that distribution.
Investment figures explain part of the gap. Median AI spend in Africa is 2% of revenue, against 5% among AI leaders. Only 32% of African firms believe their current AI investment is sufficient, against 55% among market leaders — and PwC’s earlier Africa CEO Survey put the figure even lower, at 26%.
The widest single gap in the study is on industry convergence — the use of AI to compete across, rather than within, traditional sector boundaries. Africa scores 5.8 on this measure against 7.1 for AI leaders. PwC argues this matters disproportionately on the continent, where many of the largest growth opportunities sit at the intersection of sectors: banking and telecoms in financial inclusion, energy and mining in power access, and logistics and finance in agriculture.
Workforce data complicates the picture, though with an important caveat — surveys of this kind tend to skew toward office-based, connected workers and are not representative of the broader African workforce. With that in mind, PwC’s Africa Workforce Hopes and Fears Survey found that 64% of African workers have used AI at work in the past 12 months, against a global average of 54%. Some 76% said generative AI improves the quality of their work. Yet only 36% of African organizations say their employees trust AI-generated insights enough to act on them, against 60% among leaders.
The report attributes part of the execution gap to operating context. “African business leaders have learnt to make strategic decisions under constraints,” PwC said. Years of macroeconomic and regulatory turbulence, the firm argues, have produced a corporate culture that prizes stability over reinvention — a posture that may now be holding back AI-driven growth.
PwC’s prescription — spend more, scale faster — comes with risks of its own. Jason Harrison, chief operating officer of creative company Up&Up Group, said in a recent interview that there is a wide gap between the promises of AI’s transformative capabilities and the reality on the ground, with many organizations ramping up capital expenditure only to see no return on investment. His preferred approach is the opposite of PwC’s: smaller experiments to determine which projects are worth scaling, without committing the full budget upfront.
Harrison also questioned the premise of the global benchmark itself. “The tricky thing is that [the AI hype cycle] is being led by America and companies in America are using that to push up their multiples for their stakeholders and their board. Companies don’t trade at those multiples normally, so people are buying the hype. Silicon Valley is also pushing the hype,” he said. “When you talk about technology strategy and a company’s technology stack, the conversation has swung too much towards AI.”
