75% of C-suite leaders say agentic AI lives up to the hype or is underestimated. 48% still plan to cut headcount.

Three-quarters of C-suite leaders at some of the world’s largest companies say agentic AI either lives up to the hype or is being underestimated. That is according to a poll conducted by AI Infra Summit at an exclusive CEO event in April, where attendees represented firms including Amazon, Dell Technologies, FedEx, Hitachi, Lenovo, MasterCard, Mercedes-Benz, Wayfair, and Zoom.

The headline figure is bullish: 50% said the hype around agentic AI is justified, 25% said it is under-hyped, and only 25% called it over-hyped. But the data on what that enthusiasm means for jobs is more ambiguous, and the survey’s sample of 29 respondents is small enough that individual answers shift the percentages by three points.

The jobs split

On headcount over the next three years, executives were divided almost evenly. 52% said teams would either grow (14%) or produce more at the same size (38%). The remaining 48% expect reductions: 10% said they would replace a significant number of roles with AI agents, while 38% anticipate shrinking headcount through AI automation.

The results sit within a broader pattern. A Mercer survey of 825 C-suite leaders found that virtually all expect at least some AI-driven headcount reduction in the next two years. A Chief Executive survey found a third of companies already using AI to cut roles, even as 53% plan workforce expansion. Fortune reported that 66% of CEOs are freezing hiring while increasing AI spending. The signals are consistently mixed: companies are bullish on AI and cautious about people, simultaneously.

Ed Nelson, co-founder of AI Infra Summit, said the conversation among attendees had moved past whether AI works. “The consensus was that high-level discussions have moved on from whether AI works to understanding how agents can be used effectively,” he said. “Some leaders were discussing how AI has already saved their organisations hundreds of millions of dollars.”

Beyond the pilot phase

The survey found that 95% of respondents said AI has moved beyond the pilot stage in their organisations. 40% said AI is already embedded in core products and strategy. These figures are consistent with broader industry data: Google reported at Cloud Next 2026 that 89% of business teams are already using AI agents, with the average organisation running 12.

On budgets, 36% said AI is cannibalising traditional IT spend, with organisations cutting regular IT budgets to fund AI. Another 46% said they are finding new, additive budgets specifically for AI. The hybrid approach to models dominated too: 85% use a mix of third-party foundation models from providers like Microsoft, OpenAI, and Google alongside proprietary layers built on top.

Generalists over specialists

The executives discussed which roles would grow as AI agents take on more work. The consensus, according to Nelson, was that generalists will prevail: people with lateral thinking, analytical ability, and the capacity to make connections across domains. Deep sector specialisms, by contrast, are likely to be eroded as AI agents absorb domain-specific knowledge.

“Integrating AI is a human resources issue, but it shouldn’t be framed as simply an upskilling and retraining exercise,” Nelson said. “It is a massive operational challenge. This isn’t about sprinkling AI on top of poor processes.”

The caveats

The survey’s sample size, 29 respondents, is too small to be statistically representative of Fortune 500 or 1000 companies. The results should be read as a directional snapshot of sentiment among a self-selected group of senior AI decision-makers, not as a reliable measure of enterprise-wide behaviour.

The findings also come from an event organised by a company that sells tickets to AI infrastructure conferences. The framing is inherently optimistic. Larger surveys from PwC, Grant Thornton, and Mercer, with sample sizes in the hundreds or thousands, paint a more complicated picture. Writer’s 2026 enterprise AI report found that 79% of organisations face challenges adopting AI, a double-digit increase from 2025, and 54% of C-suite executives admitted AI adoption is “tearing their company apart.”

The gap between executive enthusiasm and organisational reality remains the defining feature of enterprise AI in 2026. Leaders say AI works. They also say their organisations are not built to use it. Those two statements are not contradictory, but they do explain why the jobs question remains unanswered even as the technology matures.