A new global study has put a number on something most business leaders quietly sense: the gap between believing you are ready for AI and actually deploying it at scale is enormous.

Infor’s Enterprise AI Adoption Impact Index, released April 22, surveyed 1,000 C-suite, VP, Director, and head-of-department professionals across the United States, United Kingdom, Germany, and France. The industries covered span retail, food and beverage, industrial manufacturing, automotive, and logistics. The research was conducted between March 24 and April 9, 2026 — making it one of the freshest reads on enterprise AI sentiment available right now.

The headline finding is striking: 80% of respondents believe their organisation has the internal capability to manage an AI implementation. Yet in practice, 49% are still in the earliest stages — running limited pilots, having paused their programs, or not yet started at all.

That is a 31-point confidence-execution gap, and it has real consequences.

Why Businesses Are Stalling

When respondents were asked to name the single greatest barrier to advancing their AI strategy, three problems dominated:

1. Data security, sovereignty, and compliance — 36%. Concerns about where data lives, who can access it, and whether AI systems meet regulatory requirements remain the number-one blocker. With the EU AI Act entering full enforcement in August 2026 and state-level rules multiplying across the US, this is not a worry that fades on its own.

2. Lack of internal AI talent — 25%. One in four businesses say they simply do not have the people to configure, maintain, and improve AI systems once they are in place. Buying a platform is easy; knowing what to do with it is another matter entirely.

3. Unclear business benefits or return on investment — 23%. Nearly a quarter of decision-makers cannot yet articulate why their organisation should push past the pilot phase. They have experimented, but the numbers justifying a full deployment have not materialised.

Together, these three barriers account for more than 80% of the stalling. They are not primarily technical failures — they are strategic, organisational, and educational ones.

The Confidence Trap

The 80% confidence number deserves attention. It is not that business leaders are pessimistic about AI — they are broadly optimistic about their ability to handle it. The problem is that confidence in capability and actual deployment readiness are two different things.

Many organisations have run enough pilot projects to feel familiar with AI tools. But familiarity is not the same as having the governance frameworks, data infrastructure, skilled teams, and clear ROI models needed to move from controlled experiments into production-scale operation.

The study suggests a pattern where companies overestimate their readiness at the outset, hit the wall of real-world complexity — compliance questions, staffing gaps, inconclusive pilot results — and stall. The pilot phase becomes a permanent home.

What This Means for Business

If your organisation is in the 49%, there are three practical things to do.

First, separate the compliance question from the AI strategy question. Data security concerns are legitimate, but they should not become an indefinite blocker. Working with advisors who understand both AI architecture and regulatory requirements can help you build guardrails without building walls. This is the kind of work a fractional AI advisor or structured AI strategy engagement is designed to handle.

Second, invest in building internal AI literacy before you invest in more tools. The talent gap is rarely about hiring AI PhDs. It is about having managers, analysts, and operations leads who understand what AI can do in their specific function, how to evaluate outputs, and how to course-correct when something goes wrong. That is a training and education challenge, not a recruitment one.

Third, define what a successful full deployment actually looks like before you scale. Many pilots stall not because they failed but because no one defined success clearly enough to make the case for the next stage. If your pilot cannot tell a clear cost or revenue story, that is a measurement problem to fix before scaling, not after.

A Moment of Accountability for AI Leadership

The Infor data arrives at a moment when the broader narrative around enterprise AI is triumphalist. Billion-dollar infrastructure deals, model launches every few weeks, and analyst forecasts projecting trillions in value creation. Against that backdrop, a study showing that half of businesses are still stuck in pilot purgatory is a useful counterweight.

The technology is available. The willingness is there. What is missing, for nearly half the business world, is the bridge from intent to execution — and that bridge is built from governance, talent, and clear strategic thinking, not more software.

---

If you’re building with Claude or Codex right now, grab the free Working With Claude field guide. Thirty-two pages on the full ecosystem, Claude Code in depth, and how to roll agents out properly. Get the free guide.