According to the Applied AI Index 2026, published by Boston Consulting Group on Wednesday, telecommunications operators have clearer strategies for using artificial intelligence (AI) than any other industry, yet they demonstrate some of the lowest gains from AI implementation.
In the report, which covers 10 sectors (without detailing South Africa or Africa), telecommunications ranked first in strategic clarity, scoring 53 out of 100 points. This figure surpassed those of the technology sector (50) and banking (45). Nevertheless, telecom companies reported an average revenue increase from AI of only 0.6%, the lowest figure alongside insurance, while cost reduction was 1.1%.
Technology companies achieved a revenue increase of 2% and a cost reduction of 2.1%. Meanwhile, banks spend the largest share of funds among the represented sectors, directing 5.7% of annual revenue to AI, compared to 4.1% for technology companies and 3.7% for telecom companies. These investments allow banks to achieve a revenue growth of 1.2% and a cost reduction of 1.6%.
A year ago, the BCG 2025 study showed that only 5% of companies were 'built for the future' and generating significant value from AI, while 35% were in the scaling phase, and 60% derived minimal benefits. Currently, the share of such 'future-oriented' companies has grown to 7.5%, and another 41% are actively scaling AI and outperforming competitors. Together, these groups account for nearly half of the 1,330 senior executives surveyed by BCG.
Future-oriented companies delivered shareholder returns 2.3 times higher than laggards, and also demonstrated revenue growth 2.4 times and EBITDA growth 2.8 times.
Nicolas De Belfond, BCG's global AI leader and co-author of the report, noted: 'The perception is that AI is not delivering results. Our data shows that the picture is more nuanced: almost half of companies are already creating real value from it, while many others are still trying to translate investments into impact.'
IT Spending Category
BCG estimates AI spending at 3.3% of revenue, which is double the 1.7% that executives participating in the AI Radar survey expected to spend in 2026. It should be noted that two surveys used different samples. Over 80% of current AI spending is now outside corporate IT budgets, specifically in areas related to AI in products, talent, and governance.
Michael Greb, Managing Director and Senior Partner at BCG, stated: 'Companies that still view AI solely as an IT spending category are underestimating both their expenditures and the potential return on them.'
Standard Bank can serve as an example: 72% of its employees are active users of generative AI, yet its overall IT spend in the banking sector has only increased by 2% in the six months leading up to June, with software and cloud service costs rising by 6%.
Gartner forecasts that IT spending in South Africa will grow by 19.8% this year to $28.1 billion, significantly outpacing the global average of 14.2%. However, BCG believes that such IT forecasts may underestimate total AI spending.
A more serious issue, according to BCG, is governance. Approximately 42% of companies expect AI agents to operate autonomously by 2030, but only 5% have implemented all six types of controls defined by the consulting firm. These controls include rollback gateways, audit trails, and agent memory processing rules.
Jeff Walters, Managing Director and Senior Partner at BCG, emphasized: 'This gap is the defining challenge for the next two years. It will not be solved by regulators or technology providers. The responsibility lies with the companies implementing these systems.'
This places responsibility on local companies. In South Africa, there is no specific AI law, and the government is revising its national AI policy with a lighter, industry-specific approach after the first project was withdrawn in April due to fabricated quotes.
Governance pays off. BCG argues that having controls adds value rather than slowing down companies: those who have implemented all six types of controls within the enterprise receive approximately three times more value from agentic AI than those with only one. The share of agentic AI in total AI value grew to 22% this year from 17% in 2025, and BCG predicts it will reach 39% by 2030.
Agents are already working at the local level. Lindelani Ramukumba, IT Director at Absa Business Banking, wrote in TechCentral in April that a multilingual support agent resolves about 40% of requests without human intervention.
Surveying companies expect to reduce the workforce by approximately 10%–15% by 2030, with reductions concentrated in middle management and front-line leadership. The results of the BCG AI at Work study in South Africa for June showed that 79% of local office workers use AI regularly, compared to 74% globally.
Amanda Luther, Managing Director and Senior Partner at BCG, noted: 'How you reorganize work and retrain your workforce matters far more than the number of jobs cut.'
These findings come with caveats. AI maturity and value metrics are self-reported, which BCG admits can cause perceptual bias, and more than half of the surveyed companies have annual revenues exceeding $5 billion. Respondents represented over 60 countries in the Asia-Pacific region, Europe, and North America. Furthermore, BCG sells AI transformation services that it recommends in its research. — © 2026 NewsCentral Media