Senior leaders are not struggling to see AI’s potential. They are struggling to convert that potential into repeatable organizational performance. The research is increasingly consistent on this point: AI use is broadening quickly, but enterprise-level scale, impact, and workforce adoption remain uneven. McKinsey’s 2025 global survey found that 88% of respondents say their organizations regularly use AI in at least one business function, yet only about one-third say their companies have begun to scale AI across the enterprise. Deloitte’s 2024 year-end generative AI findings echo the same pattern: leaders remain bullish, but most expect only a minority of their experiments to scale in the near term.
That gap matters because it changes the leadership task. The challenge is no longer whether the C-suite believes AI matters. At most large organizations, that debate is over. The challenge is whether senior leaders can steer AI adoption as an operating transformation rather than a technology initiative. McKinsey’s 2025 workplace report makes that point directly: the biggest barrier to scaling AI is not that employees are unwilling, but that leaders are not steering fast enough or clearly enough.
They Are Confusing Activity with Adoption
One of the most common executive mistakes is to interpret visible AI activity as proof of meaningful adoption. A few pilots, some successful demos, a cluster of use cases in marketing or IT, and a growing stack of vendor relationships can create the impression that the organization is progressing quickly. But this is often adoption in the shallow sense—tool use without operational reinvention. McKinsey’s 2025 AI survey shows that while AI use is widespread, most organizations still have not embedded it deeply enough in workflows and processes to generate material enterprise-level impact. Only 39% of respondents reported EBIT impact at the enterprise level, even as use-case-level benefits were more common.
This is where senior leaders often overestimate progress. AI appears to be “in the business,” but it is not yet changing how the business works. That distinction is crucial. The companies getting more value from AI are not simply adding tools; they are redesigning workflows, clarifying accountability, and integrating AI into how decisions and tasks actually move. McKinsey found that high performers are nearly three times more likely than others to have fundamentally redesigned workflows, while BCG’s 2025 survey similarly points to companies moving beyond productivity plays toward end-to-end workflow reshaping.
They Are Underestimating the Human Side of Adoption
Senior leaders often speak about AI as though it were primarily a technology implementation issue. The research suggests otherwise. IBM’s 2024 global CEO study found that 64% of CEOs say success with generative AI will depend more on people’s adoption than on the technology itself, while 61% admit they are pushing their organizations to adopt generative AI faster than some employees are comfortable with. In other words, many CEOs understand that people are central to adoption even as they are creating the conditions for resistance.
This tension becomes more visible further down the organization. BCG’s 2025 AI-at-work survey found a pronounced usage gap between leaders and frontline employees: more than three-quarters of leaders and managers say they use generative AI several times a week, but regular use among frontline employees is only 51%. The same research shows that leadership support matters sharply; the share of employees who feel positive about generative AI rises from 15% to 55% when leadership support is strong. Yet only about one-quarter of frontline employees say they receive that support.
The practical implication is that many senior leaders are trying to scale AI without translating it into a workforce experience people can actually absorb. Employees are not just being asked to use new tools. They are being asked to change how they work, how they make decisions, and in some cases how they understand the future of their roles. BCG also found that employees in organizations undergoing deeper AI-driven redesign are more worried about job security than those in less advanced firms, suggesting that workforce concerns remain unresolved even in companies moving faster.
They Are Pursuing ROI Without a Coherent Change Model
The pressure for measurable returns is growing, and understandably so. BCG’s 2025 AI Radar findings show that three-quarters of executives rank AI as a top-three strategic priority, and Deloitte reports that almost all organizations with their most advanced generative AI initiatives are seeing measurable ROI. But both firms point to a more nuanced reality: the issue is not whether ROI is possible, but whether organizations are building the management systems needed to reach it consistently. Deloitte notes that business adoption is moving at the speed of organizational change, not the speed of the technology itself.
This is where many senior leaders struggle. They want financial results quickly, but they often underinvest in the slower work that makes those results durable: redesigning workflows, changing governance, reskilling teams, redefining roles, and setting clearer decision rights. McKinsey’s findings are instructive here. High performers are more likely to combine efficiency goals with growth and innovation goals, to redesign how work is done, and to have senior leaders who demonstrate strong ownership and commitment to AI initiatives. That suggests that AI value is less about isolated use cases and more about whether leaders can orchestrate enterprise change around them.
In practice, this means many executive teams are stuck in an awkward middle ground. They have moved past the initial hype, but they have not yet built a robust operating model for AI. They want ROI from a portfolio of experiments, but have not decided which ones deserve scale, how scale will be governed, or how business units will absorb the change. The result is not a lack of innovation. It is a lack of managerial conversion.
They Are Treating Governance as a Delay Rather Than a Design Choice
Another recurring struggle is governance. Leaders often understand that regulatory, legal, model-risk, and security concerns matter, but they still behave as though governance can be layered on after the fact. Research suggests that this is exactly where many organizations are getting stuck. Deloitte’s year-end 2024 report found that regulatory compliance had become the top barrier holding organizations back from developing and deploying generative AI tools, and that 69% of respondents expect it will take more than a year to fully implement a governance strategy.
McKinsey’s 2025 survey points in the same direction. Among organizations using AI, 51% report experiencing at least one negative consequence from AI use, with inaccuracies cited by nearly one-third of all respondents. High performers are more likely to have defined processes for when model outputs need human validation and more likely to have strong leadership ownership in place. In other words, governance is not just a compliance burden; it is part of what differentiates organizations that can scale responsibly from those that remain trapped in fragmented experimentation.
This is a C-suite challenge because governance is not purely technical. It requires cross-functional clarity about acceptable risk, escalation paths, validation standards, and who owns the decision when an AI-enabled process affects customers, employees, or core operations. Senior leaders often struggle here because they are accustomed to separating innovation from control functions. AI makes that separation harder to sustain. If governance is weak, scale slows. If governance is overbuilt and disconnected from the business, scale also slows. The real work is designing governance that is rigorous enough to protect trust without making forward movement impossible.
They Have Not Solved the Capability Problem
Finally, senior leaders continue to underestimate the degree to which AI adoption is a capability-building problem. It is not enough to announce strategy, buy tools, and encourage experimentation. People need time, training, and support that are specific enough to change day-to-day behavior. BCG’s 2025 employee survey found that regular usage is sharply higher among employees who receive at least five hours of training and have access to in-person training and coaching, yet only one-third say they have been properly trained.
This aligns with the broader pattern in the research: the organizations seeing more value from AI are not relying solely on enthusiasm. They are investing in talent strategies, workflow redesign, and leadership ownership. McKinsey found that the practices most associated with AI value span strategy, talent, operating model, technology, data, and adoption. That is a useful corrective to the idea that AI adoption is mainly about deploying models or tools. At scale, it is about building an organization capable of working differently.
For senior leaders, that means the real challenge is less about selecting the right pilot and more about creating the right system. They need to decide where AI should genuinely reshape work, where human judgment must remain central, how the organization will train people in role-specific ways, and what success should look like beyond experimentation. The research increasingly suggests that adoption stalls not because the technology is underpowered, but because leadership systems are underdesigned.
The Leadership Standard Is Changing
AI adoption is now exposing a broader truth about executive leadership. This is not simply a test of whether the C-suite can sponsor innovation. It is a test of whether senior leaders can drive organizational change with enough clarity, discipline, and patience to turn experimentation into institutional capability. The companies that succeed are not necessarily the ones moving fastest in public. They are the ones whose leaders can align strategy, talent, governance, workflows, and incentives around a coherent operating shift.
That is why the most important question for senior leaders is no longer, “Are we using AI?” It is, “What in our organization must change for AI to become real enterprise value?” Until more executive teams answer that question rigorously, adoption will keep looking broader than it feels—and much broader than it proves.
