Nearly seven in ten chief executive officers say artificial intelligence has already begun transforming the parts of their business they consider core, but the technology's promised leap from efficiency tool to growth engine remains elusive, according to the IBM Institute for Business Value's 2026 CEO Study.

The annual survey, which polled 2,000 CEOs across 33 geographies and 21 industries between February and April this year in partnership with Oxford Economics, found that only 10% of respondents said advanced AI or agentic AI was primarily driving growth. That is a sharp fall from the 49% who, two years ago, expected generative AI to be their primary growth engine by 2026.

The gap between expectation and reality has pushed productivity and profitability to the top of the CEO priority list for the first time, displacing AI and technology modernisation, which now sits at number two.

Must read: From layoffs to leadership shake-ups: How AI reshaped tech in April 2026

The gap between hype and delivery 

The data reveals what the report calls a structural gap between CEO optimism and enterprise readiness. In 2024, nearly half of all CEOs surveyed predicted that advanced AI would primarily drive growth by 2026. By 2025, that forecast had moderated to 15%. The actual 2026 number came in at 10%.

Instead of growth, the dominant mode of AI investment this year is still piloting and experimenting, cited by 53% of CEOs. Another 30% say efficiency and savings is their primary objective, down from 51% in the 2025 prediction for the same year.

One of the study's more striking data points is the surge in Chief AI Officer (CAIO) appointments. In 2026, 76% of organisations say they have someone in this position, compared to just 26% in 2025. All of those CEOs expect the CAIO's influence to increase further by 2030.

The report argues that this role, when given real authority over AI priorities, funding, and standards, rather than ownership of business outcomes, creates the conditions for scaled transformation. It positions the CAIO as an accelerator of decisions, not an owner of results.

Must read: AI layoffs may hurt companies too, not just workers: Study warns of ‘automation trap’

At the same time, 85% of CEOs say all functional leaders must now become technology experts in their domains, a finding that points to the broader restructuring of C-suite roles that the study predicts will be complete by 2030.

Decisions moving to machines

The study finds that AI is already operating autonomously in a significant share of business decisions. CEOs say 25% of operational decisions are currently made by AI without human intervention, typically in areas such as pricing updates, inventory allocation, shipment rerouting, and automated incident remediation. By 2030, CEOs expect that share to nearly double to 48%.

Confidence in AI-generated inputs for strategic decisions is also growing. Sixty-four percent of CEOs say they are comfortable making major strategic decisions based on AI-generated input. This marks a notable change from 2025, when 62% said generative AI was still too risky for core business functions.

A central argument of this year's report is that competitive advantage will come not from which AI model a company uses, but from how deeply it is customised with proprietary data and intellectual property. CEOs who embed their own data into custom AI models and agents expect 13% more of their 2030 revenue to come from products and services that do not exist today, the study notes.

Currently, 39% of CEOs say their organisation primarily uses pre-trained foundation models. By 2030, that is expected to fall to 13%. Half of all CEOs say they are shifting to a hybrid strategy that blends custom models, foundation models, and smaller specialised models, depending on specific business needs.

The report also finds that 83% of all CEOs, and 97%  of those classified as AI-first, say developing and maintaining AI sovereignty is essential to their business strategy.

The human side of the equation

Despite widespread belief in AI's potential, the study finds a significant deployment gap. CEOs estimate that only 25% of their workforce uses AI regularly as part of their job. Yet 86% of the same CEOs say their employees have the skills to collaborate with AI. The report concludes that this is an organisational design problem, not a skills problem.

The reskilling numbers are substantial. CEOs say 19% of their workforce was reskilled in the past year to perform a different role, while 41% were reskilled to perform their current role more effectively. Between 2026 and 2028, CEOs expect 29% of employees to require reskilling for a different role.

Employee sentiment on AI is broadly positive. Across all generational groups, at least twice as many workers say they would embrace greater AI use rather than resist it. Sixty-one percent say AI makes their job less repetitive and more strategic.

The payoff from redesigning how human and machine intelligence work together is measurable. CEOs who have redesigned all five core business areas, technology, finance, HR, operations, and cross-functional collaboration, are four times more likely to have fully delivered on the business cases they set out to realise.