Why agentic AI is stuck in first gear?
Corporate AI spending is rising sharply, with 21% of leaders investing over $10 million. Yet scaling agentic AI is hindered by data gaps, workforce fears, and governance risks.
Corporate AI spending is rising sharply, with 21% of leaders investing over $10 million. Yet scaling agentic AI is hindered by data gaps, workforce fears, and governance risks.
Artificial intelligence has moved from the margins to the center of corporate strategy. Across industries, companies are increasing their financial commitments, yet many still struggle to realize AI’s full potential.
EY’s latest AI Pulse Survey paints a complex picture: enthusiasm and investment are rising, but execution—particularly for agentic AI—lags behind.
The findings carry significant implications for CFOs charged with balancing ambition, governance, and risk.
The pace of AI investment is accelerating. According to the survey, 21% of senior leaders say their organizations have invested more than $10 million in AI this year, up from 16% in 2024. That figure is expected to climb to 35% in 2026, signaling a belief that AI will remain a priority in corporate budgets.
For many companies, those bets are paying off. 97% of respondents report positive ROI across AI projects, with the strongest results coming from organizations that dedicate at least five percent of their budget to the technology.
Investments are increasingly targeted toward building custom AI solutions in-house rather than relying solely on acquisitions—a sign of growing confidence in internal capabilities but also a risk, given the cost and complexity of developing bespoke systems.
While general AI adoption is delivering returns, agentic AI—systems that can act autonomously to achieve complex goals—remains in early stages.
Only 14% of organizations report full deployment, even though 34% have begun implementation. The reasons for slow progress are clear: 87% of leaders identify significant barriers, including cybersecurity and data privacy concerns, a lack of regulatory clarity, and the absence of internal policies to govern these systems.
The survey also uncovers a psychological hurdle. 64% of senior leaders believe that fear of AI replacing human jobs will hold back adoption. This anxiety mirrors a broader cultural resistance, where leaders acknowledge AI’s capabilities but remain cautious about its implications for the workforce.
As Dan Diasio, EY’s Global Artificial Intelligence Consulting Leader, observed, companies see agentic AI as both impressive and intimidating, forcing executives to rethink how technology and human capital interact.
Adoption challenges are not confined to technology. Governance and data readiness are emerging as critical battlegrounds. 70% of respondents say poor data quality is one of the biggest obstacles to scaling AI, reinforcing the need for robust data frameworks and governance.
Without them, the risk of flawed insights and costly errors grows.
At the same time, companies are paying more attention to ethics. Training programs on responsible AI have expanded significantly, with 59& of organizations increasing efforts this year and 64% planning to do so in 2026.
The emphasis is shifting from ad hoc compliance to embedding ethical considerations into AI design from the outset. This focus reflects a recognition that governance is not optional: without trust, adoption will stall, no matter how much money is spent.
These findings align with shifts in corporate oversight identified in EY’s 2025 proxy season review, where boards are taking a more active role in technology governance.
The number of S&P 500 companies assigning AI oversight to board committees has more than tripled this year, and nearly half of Fortune 100 companies now highlight AI expertise in director qualifications.
Investors are watching closely, signaling that companies must not only deploy AI effectively but also demonstrate they are managing its risks.
For CFOs, AI is no longer a side project. It is a capital allocation decision with implications for growth, risk, and stakeholder confidence.
The survey suggests that success will depend on three factors:
Companies that build strong data foundations, invest in responsible AI, and maintain clear oversight are positioned to convert today’s incremental gains into enterprise-scale value. Those that do not may find themselves spending heavily without securing the returns investors expect.