STUDY: A New IBM Study Of 2,000 Technology Executives Across 33 Countries Finds That Two-Thirds Of CIOs And CTOs Are Now Formally Accountable For AI Systems They Do Not Fully Control, With Organizations Averaging 54 AI-Related Incidents Per Year 🤖

The IBM Institute for Business Value, working with Oxford Economics, surveyed 2,000 C-level technology executives across 33 geographies and 19 industries between January and April 2026, making it one of the largest and most geographically comprehensive studies of enterprise AI governance ever conducted. The central finding is structural rather than attitudinal: 70 percent of surveyed executives say teams across their organizations are deploying AI technology faster than IT can actually track, meaning the people formally responsible for their company’s AI systems frequently do not know what is running, where it is running, or what it is doing. This is not a complaint about the pace of innovation. It is a governance failure of a specific and dangerous kind: accountability without visibility, where the executive who is responsible if something goes wrong does not have the information necessary to prevent it from going wrong in the first place. The average organization in the study is now experiencing 54 AI-related incidents per year, of which 37 percent resulted in data exposure or security breaches, numbers that would generate significant regulatory scrutiny in any other operational domain but are currently treated largely as acceptable costs of the deployment pace.

The financial management picture is as broken as the governance picture. Eighty-four percent of technology executives have not fully operationalized AI financial management, meaning they do not have reliable systems for tracking what AI is actually costing them across the enterprise, and 85 percent lack full visibility into real-time AI spend, a fact that becomes particularly consequential given that AI agent deployments are projected to increase by 38 percent by 2027 and that the cost of running large-scale agentic AI systems scales directly with usage in ways that traditional software licensing did not. Yet the study also identifies a clear structural divide between organizations that have built control into their AI systems from the architecture level and those relying on manual governance after the fact. Organizations in the first category deploy 16 times more AI agents than those relying on manual governance, deliver 18 percent higher operating margins, spend four times less of their AI budget, and experience 25 percent fewer incidents. The performance gap is not marginal and it does not appear to be coincidental: it reflects the compounding advantage of knowing what your systems are doing versus discovering it after something breaks. IBM’s own CIO Matt Lyteson framed the challenge directly, saying it is no longer just about deploying AI faster but about redesigning how organizations control, govern, and invest in it, embedding control and visibility from the start so they can scale with confidence.

The 11 percent preparedness figure is the most telling single data point in the entire study. Eighty percent of surveyed executives report operating under a CEO-driven mandate to accelerate AI transformation, meaning the pressure to deploy faster is coming from the top of the organization and is not optional. Yet only 11 percent believe they are fully ready for the scale of AI agent deployment expected within the next 12 months, and 77 percent report that their current governance capabilities are already falling behind the pace of adoption. That combination, a mandatory acceleration with almost no organizational readiness and documented governance failure already underway, is not a planning problem that will resolve itself as organizations gain more experience with the technology. It is a structural conflict between the incentive structure driving deployment decisions, which rewards speed and scale, and the risk management requirements of operating autonomous AI systems across enterprise environments, which require visibility, control, and accountability structures that take time to build and that no one is currently being rewarded for prioritizing. Security and compliance concerns are cited as the top barrier to scaling AI agents by 59 percent of executives, a majority who simultaneously report being unable to stop or slow deployment because the mandate to accelerate is coming from their CEO.