IBM Study: Cultural Gaps, Not Tech Limits, Block AI Value

Organizations racing to deploy artificial intelligence are discovering their biggest obstacle isn't computational power or algorithmic sophistication—it's the widening gap between how leadership perceives AI adoption and how employees actually experience it.

A global study from the IBM Institute for Business Value reveals a striking disconnect: executives report AI-driven role changes at twice the rate employees observe in their own work. While 81% of executives believe workers are rewarded for building AI capabilities, 43% of employees say their employer provides no AI training whatsoever.

The findings, shared by Aparna Nair, IBM's Chief Talent, Leadership & Culture Officer, point to a fundamental misalignment in how companies approach AI transformation. Nearly two-thirds of surveyed executives acknowledge AI is reshaping workflows, yet most organizations haven't redesigned the management practices, performance systems, and support structures required to make those changes effective.

Why it matters

This perception gap directly threatens return on investment. When employees lack training, clear accountability frameworks, and psychological safety to question AI outputs, even sophisticated technology deployments fail to deliver value. The research shows organizations that address these human factors achieve 73% higher revenue growth and an 11-percentage-point operating margin advantage over peers—demonstrating that cultural readiness drives financial outcomes as much as technical capability.

Trust and psychological safety emerge as critical barriers

The study, titled "Where AI breaks—or breaks through," identifies a troubling pattern: 43% of executives say employees don't feel safe raising concerns about AI outputs, and more than half of employees report colleagues fail to challenge AI recommendations when they should. This reluctance to question automated systems creates risk, particularly as AI takes on higher-stakes decision support roles.

The issue extends beyond individual comfort. When organizational culture discourages critical thinking about AI results, the problem isn't the tool itself but how companies enable or discourage its responsible use.

Managers face new demands without adequate support

The transition places unprecedented pressure on middle management. Managers must now coach judgment, oversee human-AI collaboration, and evaluate performance using criteria that didn't exist until recently. Yet 93% of executives acknowledge AI-enabled work has made performance evaluation significantly harder, while only about a quarter of employees say their managers focus primarily on coaching and strategic judgment—a proportion expected to nearly double by 2028.

This gap suggests most organizations lack the infrastructure to support managers in their evolving roles. The shift from directing tasks to developing decision-making capabilities across teams requires new skills, clearer expectations, and purpose-built tools.

What distinguishes high performers

A small subset of organizations have achieved both advanced AI maturity and strong change management capabilities. These leaders share common practices: they define clear accountability, establish explicit norms for when to trust or challenge AI, and align performance systems to reflect how work has actually changed.

Nair recommends three priorities for people leaders: treat AI workflow changes as adoption challenges requiring defined skills and behaviors before scaling; update performance systems to reward learning and sound judgment, not just speed; and create practice opportunities through simulations and real-world examples that teach employees when to trust, challenge, or escalate AI decisions.

The research was first reported by IBM in a blog post by Nair and detailed in the full IBM Institute for Business Value study.