Deloitte: 79% of IT Leaders Struggle to Convert Investments in AI into Real Returns

A groundbreaking study by Deloitte involving 660 senior technology leaders globally reveals a contradiction that should concern any board of directors: 81% of CIOs express confidence in their ability to scale AI within their organisations, while 42% report low or non-existent returns on previous investments. The gap between confidence and results is not a technological problem. It is a leadership and operational model issue.

The Mandate Shift

The 2026 Global Technology Leadership Study by Deloitte documents a transformation that was already anticipated but is now measurable: 79% of technology leaders identify "generating business results" as their primary responsibility, a decisive departure from the historical focus on infrastructure reliability and systems maintenance.

Anjali Shaikh, Executive Director of Deloitte Consulting, is direct in her assessment: "The era of the operational technologist is over." What the study describes is a redistribution of expectations: the modern CIO is responsible for redesigning the company’s fundamental operations, not merely keeping systems running.

This change has an immediate cause: artificial intelligence. More than any previous technology, AI requires the technology leader to operate at the centre of strategic decision-making, because choices regarding what to automate, where to apply AI, and how to govern autonomous systems are business decisions with long-term consequences, not delegable technical choices.

The Confidence Paradox

The study’s figures reveal a concerning collective illusion. While 81% of leaders express confidence in scaling AI, 75% simultaneously recognise that their operational models need fundamental transformation to generate more value. The coexistence of these two perceptions is not contradictory, it is revealing.

Confidence pertains to the technology. Doubt pertains to the organisation.

And when the study delves into concrete results, the paradox deepens: only 29% of companies report significant ROI from generative AI tools. Only 23% see measurable returns from AI agents. Seventy-three percent of CEOs report stress regarding their company’s AI strategy, and 64% fear losing their jobs if they err in the transition.

Distributed Leadership and Coordination as a Competence

The study identifies another significant transformation: the proliferation of technical leadership. Seventy-one percent of organisations now have five or more executives with technology responsibilities. The model of the CIO as the sole point of technological control has become outdated.

Steve Pratt, Principal at Deloitte Consulting, describes the new core competency: "Today’s leaders are defined by their ability to orchestrate across the C-suite and translate technology into measurable results." Those who remain focused solely on delivering underlying systems risk marginalisation.

The Three Execution Blockers

The research identifies three structural challenges that separate high-performing AI organisations from those that lag behind:

From uptime to outcomes: technology has evolved from enabling operations to defining strategy. Leaders who still measure success by system availability rather than business impact are using the wrong metric.

Ambition versus execution capability: organisations have AI aspirations that exceed their foundation, data quality, AI-ready talent, and modern operational models. The gap between what leaders want to achieve and what their organisations can execute is the primary source of negative ROI.

Resource constraints and trade-offs: 41% of technology leaders report that business units perceive them as unable to keep pace with demand. Operating legacy systems, defending against threats, driving transformation, and accelerating growth, simultaneously, with budgets designed for another era, forces choices that compromise transformational initiatives.

What the Study Implies for Boards

The most important conclusion of the Global Technology Leadership Study is not explicitly stated in the report: organisations have promoted technology leaders to strategic roles without restructuring their organisations to support this expanded mandate. The problem is not the CIO. It is the organisational model around the CIO.

For boards of directors and CEOs, the practical implication is direct: if returns on AI investments are low, before questioning the AI strategy, consider whether the organisational structure is designed to deliver it. The difference between AI leaders and followers in 2026 will not be determined by the technology they adopt, but by the operational model they build around it.