Gartner's April 2026 CEO survey puts 80% of chief executives on record expecting AI to force operational overhauls. McKinsey's data shows only 7% of organizations have fully scaled AI. As an IS researcher, I find the gap between those two numbers more revealing than either one alone.
I was reading Gartner's April 2026 IT spending report when I hit the CEO survey figure: 80 percent of chief executives say AI will force operational capability overhauls inside their organizations. Not "will be a useful tool." Not "will improve productivity." Overhauls. That word is doing real work, and I think it is worth sitting with what it implies, because the gap between that expectation and what most organizations can actually deliver right now is significant in a way that has not gotten enough attention.
The McKinsey State of AI 2025 report sits on the other side of that expectation. While 88 percent of organizations report using AI in some form, only 7 percent have fully scaled it across their operations. Gartner's own data adds another layer: only 17 percent of organizations have deployed AI agents so far, even though more than 60 percent expect to do so within two years. So you have 80 percent of CEOs expecting operational overhauls from AI, and roughly 7 percent of organizations at a scaling level that could plausibly deliver one. The CIO sitting between those two data points is being asked to close a gap that is real and wide and mostly not talked about directly.
The gap gets misread in a particular way that I want to push back on. The common framing is that organizations are moving slowly because the technology is unproven, or because the ROI is unclear, or because security and compliance are holding things back. My read is different. Most organizations have not done the organizational groundwork that makes a genuine operational overhaul possible. The technology has raced ahead of the processes, the data infrastructure, and the workforce capacity that would need to carry it. The CIO is not the obstacle. The CIO is the person being handed a technology mandate and a timeline while the organizational prerequisites are still being built.
When a CEO says "operational capability overhaul," they are describing something concrete: the way work gets done changes fundamentally. Processes designed around human judgment get redesigned around AI-assisted or AI-directed workflows. Some roles shift. Some tasks that humans currently do get automated. New tasks appear that did not exist before. The organization becomes faster at certain things and has to develop new capabilities to govern AI systems that are now making decisions on its behalf. That is a real and achievable thing. It is also a substantial organizational change project that touches almost every function the CIO is responsible for.
The sociotechnical systems framework, developed by Trist and Bamforth (1951) in the context of coal mining reorganization, has a surprisingly direct application here. Trist and Bamforth's core finding was that technical change and social change are interdependent. You cannot optimize the technical system while treating the social system as static and expect good outcomes. The organizations that tried to introduce new mining technology without redesigning the human work structures around it got worse results than the organizations that redesigned both together. The CEO expectation of AI-driven operational overhauls is essentially a prediction about technology change. The CIO knows, or should know, that realizing that prediction requires equivalent investment in the social system: the workflows, the roles, the skills, the culture, the incentives, the accountability structures.
The CIO's structural problem is that they are being asked to deliver two things simultaneously that pull in opposite directions. The first is operational continuity. Systems cannot go down. Security posture has to stay strong, especially as AI introduces new attack surfaces and new vectors for data exposure. The infrastructure has to scale to support AI workloads, which are substantially more compute-intensive than the workloads they replace or augment. The second is transformation. Build the new system while running the old one. Move fast enough to satisfy the CEO timeline while moving carefully enough not to break what cannot break.
The CIO who is too focused on stability cannot move fast enough. The CIO who chases transformation speed too aggressively breaks things that are not allowed to break. That structural tension is not new, but AI has intensified it because the CEO expectation timeline is now explicitly tied to a technology category that is moving faster than any previous enterprise technology wave. Gartner is reporting that AI agent software spending is projected to grow from $86.4 billion in 2025 to $206.5 billion in 2026 and $376.3 billion in 2027. That is not an incremental budget line. It is a major reallocation, and it is happening while existing infrastructure still needs to be maintained and secured.
The sequencing problem is where I think the CEO expectation most often goes wrong, not in intent but in order of operations. Organizations tend to reach for technology solutions before doing the process work. They deploy an AI tool into an existing workflow and wonder why the productivity gains are smaller than the vendor projected. The workflow itself is usually part of the problem. AI tools are effective at executing workflows faster and more consistently. They are not good at fixing workflows that were poorly designed before the AI arrived. An operational overhaul requires redesigning the workflow first, then identifying where AI accelerates the redesigned process. Doing it in reverse is how you end up with an expensive technology layered on top of a process that nobody thought to challenge.
Data infrastructure sits underneath all of this and is probably the least glamorous part of the conversation. AI systems require clean, governed, accessible data. Most large organizations have data that is siloed across systems built at different times with different standards, inconsistently formatted, poorly documented, and governed by access controls designed for a world where only humans queried it. The AI pilot that looked impressive in a controlled demo was running on a carefully prepared dataset. The production AI system has to run on whatever the organization actually has. Closing that gap requires investment in data engineering, data governance, and data architecture that does not show up easily in a board presentation about AI transformation but is the actual foundation on which everything else depends.
The workforce dimension is where the CEO expectation gets most understated. Operational overhauls change jobs. The 80 percent CEO figure implicitly acknowledges that the people doing the operations being overhauled will need different capabilities. But workforce transformation takes years. Reskilling programs take time to design, fund, and deliver. The organizational memory of how things were done before has to be actively replaced with fluency in how things work now. Organizations that are serious about AI-driven transformation are building reskilling infrastructure now, before the technology is deployed at scale. Organizations that plan to figure out workforce change after deployment are going to find that the technology is ready before the people are.
What genuinely surprises me in this data is the gap between the confidence of the CEO expectation and the organizational reality that McKinsey's 7 percent scaling figure describes. I do not think that gap is explained by CEO naivete. I think it reflects something real about how responsibility for transformation is distributed inside organizations. The CEO has the expectation. The CIO has the accountability. The gap between those two is where the most interesting IS research questions about AI-enabled organizational change actually live.
---
claims_checked:
- "80% of CEOs say AI will force operational capability overhauls": "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
- "88% of organizations use AI; only 7% have fully scaled it": "https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai"
- "AI agent software spending: $86.4B (2025) to $206.5B (2026) to $376.3B (2027)": "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
- "49% of CIOs planning agent deployment in 12 months": "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
claims_unverified:
- "The 17% organizations deployed AI agents figure is Gartner data cited widely across multiple press releases; the specific source URL may differ from the April 2026 forecast"
- "Observations about workflow sequencing, CEO/CIO responsibility gaps, and reskilling timelines are analytical judgments, not specific sourced statistics; presented as the author's read"
sources_used:
- "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
- "https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai"
word_count: 1100
About the author
Share
More notes
Related notes