AI investments are accelerating globally, but understanding and full adoption of agentic AI remain limited, according to EY's latest US AI Pulse Survey.
 |
| Despite high ROI from AI, most businesses remain hesitant on agentic AI implementation, citing lack of understanding and strategic readiness, EY finds. Image: CH |
NEW YORK, United States — July 30, 2025:
Artificial intelligence investment continues to surge across industries, yet understanding and adoption of agentic AI—intelligent systems capable of autonomous goal-driven action—remain far behind, according to the newly released EY US AI Pulse Survey.
The survey, conducted in April 2025 among 500 senior U.S. business leaders, reveals a stark disconnect between corporate AI spending and practical integration of agentic technologies. While nearly all organizations investing in AI report strong returns on investment, only 14 percent have fully implemented agentic AI within their operations. An additional 34 percent say they have initiated limited-scale deployments, mostly in operational support areas like customer service, IT, and cybersecurity.
Despite this progress, most organizations still lack a clear understanding of what agentic AI entails or how to scale it effectively. More than half of senior leaders believe executives in their own organizations and industries fail to comprehend the real value of agentic AI. This knowledge gap is slowing implementation even as AI continues to prove its value.
The financial commitment to AI remains strong. Twenty-one percent of surveyed organizations have already invested $10 million or more in AI, up from 16 percent a year ago. Looking ahead, 35 percent expect to match or exceed that amount next year, signaling continued confidence in the technology’s potential. However, the survey also shows that tangible results are most visible among organizations that allocate at least five percent of their budgets to AI, outpacing others in areas such as technology upgrades, customer satisfaction, and cybersecurity improvements.
While agentic AI is gaining traction in select functions, widespread rollout faces significant hurdles. Senior leaders cite cybersecurity and data privacy risks as major concerns. A lack of industry regulation and the absence of internal company policies on agentic AI use also stand in the way. At the same time, the vision for autonomous systems remains bold, with 73 percent of respondents believing that entire business units could eventually be run by AI agents. Still, 89 percent emphasize that such systems must retain built-in human oversight to ensure responsible and effective use.
As companies confront the challenges of implementation, more are turning their focus inward. A growing number of organizations are building customized, in-house AI solutions—64 percent this year, up from 56 percent last year. Meanwhile, fewer companies are seeking to acquire external firms for AI capabilities, suggesting a maturing AI ecosystem where differentiation and internal capability development are prioritized.
Another key theme emerging from the survey is the need for workforce preparedness. Sixty-four percent of senior leaders say their organization will spend more time training employees to use AI responsibly in the coming year, a sharp increase from 49 percent the year before. While fears about job loss persist, most leaders see training—not resistance—as the path forward, with only 24 percent identifying employee pushback as a top barrier.
Dan Diasio, EY Global Consulting AI Leader, stated that “AI agents can revolutionize how we work and unlock possibilities that were once unimaginable.” However, he emphasized that leaders must navigate the complexity of implementation and strike the right balance between AI’s capabilities and human judgment. Whitt Butler, EY Americas Vice Chair for Consulting, echoed this, noting that the future of work will depend on how well organizations prepare their people, embed responsible governance, and align AI efforts with meaningful business outcomes.
The survey reflects a clear trend: while AI momentum is undeniable, the road to truly autonomous, agentic systems will be gradual, requiring deliberate strategy, trust, and an informed workforce.
The EY US AI Pulse Survey was conducted online from April 17 to 30, 2025, among 500 U.S.-based senior decision-makers (SVP level and above) across ten sectors including healthcare, energy, finance, technology, and government. This third wave maintains a margin of error of ±4 percentage points at a 95% confidence level, with consistent methodology used in previous waves for accurate year-over-year comparisons.