Artificial intelligence is rapidly spreading across the enterprise. Employees are using AI tools to improve their productivity, companies are launching pilots across a growing number of business functions, and executives are investing heavily in AI initiatives. Yet for most organizations, a major gap remains between experimenting with AI and making it a core operational capability.
Why is that gap proving so difficult to bridge?
“Organizations across industries have invested heavily in artificial intelligence initiatives, yet the transition from discrete, bounded projects to embedded, organization-wide capability has proven elusive for the vast majority,” said “A Field Study of AI Maturity, Governance, and the Evolving CIO Role,” a report published by Bee360, an IT management consulting company.
“Understanding why this transition is difficult, and what it actually requires, is a question of both scientific and practical urgency,” the report added. “What does it actually take to make AI a core operational capability — not just a portfolio of pilots?”
The question reflects what Bee360 calls “one of the defining challenges of enterprise management in the mid-2020s: the gap between AI experimentation and AI operationalization.”
And increasingly, bridging that gap is becoming a central responsibility of the CIO.
As part of its analysis, Bee360 cites McKinsey’s “The State of AI in 2025,” a report based on a survey of 1,993 respondents in 105 countries. The survey found that while AI adoption is becoming nearly universal, realizing value from AI at scale remains the exception rather than the rule. “For most organizations, AI use remains in pilot phases.”
“Against this backdrop, the guiding question synthesizes four distinct but interrelated constructs,” said Bee360: “The maturity of AI adoption, organizational barriers to operationalization, governance posture, and the evolving role of the CIO.”
Taken together, these four issues point to a broader question at the heart of enterprise AI: Is the primary challenge really about mastering a powerful new technology, or is it about transforming the organization itself?
The first stage of the Bee360 field study was conducted during the 2026 MIT Sloan CIO Symposium on May 18 and 19, where 20 CIO-level technology leaders were each asked four key questions. The second stage took place the following day, May 20, when five CIOs participated in a two-hour facilitated roundtable. To extend the geographic scope of the study, the survey was subsequently distributed to CIO-level technology leaders in Europe, receiving 25 responses over the following two weeks.
MIT Sloan CIO Symposium 2026 as a Research Site
The MIT Sloan CIO Symposium was an unusually well-suited research site for a field study of this nature, given its concentration of qualified respondents. Approximately 400 CIO-level technology executives from a wide range of industries and geographies attended the Symposium.
The interviews with the 20 CIOs during the MIT Symposium led to four major conclusions.
- AI is already changing how people work, but it is not yet embedded at scale. All 20 CIOs said their organizations were already adopting AI.
- The biggest barrier is organizational, not technological. The key challenge is redesigning workflows and managing organizational change.
- Governance is reactive or absent. Only 15% have established the kind of governance needed to define accountability for AI decisions and outcomes.
- The CIO role has fundamentally changed toward what the report calls “directive emergence”, that is, setting a direction for AI and continuously adapting their approach, rather than executing against a fixed plan.
CIO Insight Session
The day following the conclusion of the MIT Sloan CIO Symposium, five CIOs convened at the Harvard Club of Boston for a two-hour facilitated discussion under Chatham House Rules. Bee360 was free to use information from the discussion but could not reveal who made any particular comment.
The five participating CIOs shared their current challenges in making AI a core operational capability. They then discussed whether they agreed with the overall takeaways from the 20 CIOs interviewed during the Symposium and concluded by offering their own views on what it takes to make AI a core operational capability.
Four overarching themes emerged from the facilitated discussion.
- The business case problem is universal. Every industry context requires translating AI’s potential into financial language that CFOs and boards can act on.
- The mindset and culture challenge appears in every industry. Its specific form differs considerably, but the underlying organizational challenge is widespread.
- The skill and capability gap is pervasive. The particular skills required differ substantially among industries, but the need to develop new human capabilities alongside AI is common to all.
- Many questions regarding governance and accountability remain unresolved. Among the most important are: Who owns the risk? Who approves AI investment? Who defines success in AI operationalization?
US and European Survey Findings
The Bee360 report also included a detailed comparison of the US and European survey findings. Each CIO was asked for their opinion on four substantive questions.
1. AI Maturity
This first question addressed the progression of generative AI adoption from individual productivity use cases to task-embedded applications and, ultimately, process-level transformation:
Which of the following best describes where your organization stands with AI today?
None of the US respondents said their organizations had not yet embedded AI into regular day-to-day work, compared with 24% of European respondents. Twenty five percent of US respondents said individual employees were using AI for their own tasks without a coordinated organizational approach, compared with 44% in Europe.
The results were reversed when asked about the more advanced use of AI at their organizations. Forty five percent of US respondents said AI was changing how specific roles or teams work, with outputs reviewed by humans before use, compared with 28% of European respondents. And 30% of US respondents said AI was embedded in operational processes or products with governance structures in place, compared with only 4% of European respondents.
“The European online survey results present a markedly different AI maturity distribution,” said the report. “These results indicate a substantial maturity gap between the two samples, with European organizations predominantly concentrated at the early individual-use phase.”
2. Biggest Barrier
The second question focused on the barriers to AI operationalization:
What is currently the biggest barrier in your organization to making AI a core operational capability?
The largest group in both samples identified the same barrier: adding AI to existing processes without redesigning the work itself, cited by 40% of US respondents and 32% of European respondents.
Data that was not sufficiently usable for AI at scale was cited by 15% in the US and 24% in Europe. A lack of skills or willingness among employees to work differently because of AI was cited by 20% in both regions. The lack of a shared framework for deciding who is accountable for AI outcomes was cited by 15% in the US and 12% in Europe. And the inability to demonstrate the financial value of AI investments was cited by 10% in both regions.
The most striking result is the consistency across the two samples: despite their differences in AI maturity, both groups identified workflow redesign as the most significant barrier. This reinforces one of the study’s central findings — that the greatest obstacles to AI operationalization are organizational rather than technological.
3. AI Governance
The third question explored how organizations are redesigning decision rights and governance frameworks as AI increasingly influences decision-making:
Which of the following best describes how your organization currently approaches AI governance?
Only 15% of US respondents and 4% of European respondents said their organizations had defined who could override AI decisions and who was responsible when AI produced a wrong outcome.
By contrast, 50% of US respondents and 60% of European respondents said they had AI policies designed primarily to prevent problems rather than enable decisions. None of the US respondents said governance processes were slowing down AI initiatives without producing clear strategic benefits, compared with 12% of European respondents. And 35% of US respondents and 24% of European respondents said they did not yet have a governance structure specifically designed for AI decisions.
The results suggest that AI governance remains largely reactive. Most organizations are still focused on guardrails and preventing misuse rather than developing governance frameworks that enable AI-powered decision-making while clearly assigning accountability.
4. The Changing CIO Role
The fourth question addressed how the CIO role is evolving as organizations progress through different levels of AI transformation:
Which of the following best describes how your role as CIO has changed as a result of AI?
Twenty percent of both US and European respondents said they were focused on building AI capability step by step before committing to larger-scale deployment.
Twenty percent of US respondents and 24% of European respondents said they were spending most of their effort changing organizational mindsets around AI rather than managing the technology itself.
The largest difference concerned the “directive emergence” model. Fifty five percent of US respondents said they were setting a direction for AI and continuously adjusting their approach as circumstances and results evolved, compared with 32% of European respondents.
Finally, only 5% of US respondents said their role had not fundamentally changed and that AI was simply one more technology priority, compared with 24% of European respondents.
Overall, the comparison between the European and MIT Sloan CIO Symposium data sets highlights both significant differences and important similarities.
European CIOs lag behind on AI maturity but converge on the same primary barrier. Forty-four percent of European respondents report AI use at the individual level with no coordinated organizational approach, and only 4% have reached the stage where AI is operationally embedded with governance structures in place. Yet both respondent groups identified the same primary barrier: adding AI to existing processes without redesigning the work itself. This was cited by 40% of US respondents and 32% of European respondents, reinforcing the conclusion that the biggest challenge is organizational, not technological, across both geographies.
Reactive governance dominates more strongly in Europe, while the CIO role is more evenly distributed. Sixty percent of European respondents describe their governance posture as reactive guardrails, and 12% report that governance actively slows down AI initiatives — a dynamic entirely absent from the MIT Sloan CIO Symposium sample. On the changing CIO role, European responses are more dispersed, suggesting that no single leadership orientation has yet consolidated at this stage of AI adoption.
Taken together, the Bee360 field study paints a picture of enterprise AI at an important transition point. AI adoption itself is no longer the central challenge. The harder task is moving from individual tools and isolated pilots to redesigned workflows, organizational capabilities, effective governance, and measurable business value.
The findings also suggest that the CIO has a particularly important role to play in this transition. As AI becomes embedded in the way organizations operate and make decisions, CIOs will increasingly need to look beyond their traditional responsibility for technology infrastructure. They will have to help redesign work, build organizational AI capabilities, establish governance and accountability, develop the necessary skills, and bring employees along as their jobs evolve.
Perhaps the most important lesson from the study is also one of the simplest: making AI a core operational capability is not primarily a technology problem. It is an organizational transformation problem. And as enterprises move from experimenting with AI to operationalizing it at scale, managing that transformation may well become one of the defining responsibilities of the CIO.