AI budgets are exploding while CIOs and CTOs lose visibility into what they're actually spending.

A new IBM Institute for Business Value study of 2,000 C-level technology executives reveals a dangerous control gap: 67% of CIOs and CTOs report being held accountable for AI systems they do not fully control, while 84% have not operationalized AI financial management.

The numbers tell a stark story. AI spend is projected to grow from 15% of IT budgets in 2025 to 25% by 2027—a 71% increase in two years. Yet 85% of tech leaders still lack full visibility into real-time AI spend. As one executive put it: "It's like flying a plane at 10,000 feet, being told to climb to 12,000, replace both engines mid-flight and ensure zero turbulence."

The Governance Crisis: Speed vs. Control

The study, conducted from January to April 2026 across 33 geographies and 19 industries, reveals that 70% of surveyed executives say teams across the business are deploying technology faster than IT can track. This isn't shadow IT from 2015—this is distributed AI procurement at scale, with departments buying model access, agent platforms, and inference capacity without central visibility.

The pressure to move faster is relentless. 80% of respondents report CEO-driven AI transformation mandates, yet only 11% believe they are fully ready for the scale of AI agent deployment expected in the next year. By 2027, surveyed tech CxOs anticipate a 38% increase in the number of AI agents deployed.

Meanwhile, 77% of organizations report AI adoption is already outpacing current governance capabilities. Governance models designed for quarterly software releases are now managing systems that operate continuously and autonomously.

Security and Operational Risks at Scale

The consequences of this control gap are already visible. Surveyed organizations experienced an average of 54 AI agent incidents last year—unintended or harmful occurrences requiring human correction. Of those incidents, 17% were high severity, requiring more than four hours to contain.

The breakdown of high-severity incidents:

• 37% resulted in data exposure or security breaches
• 33% caused cascading system failures
• 17% triggered compliance issues

Victoria Medina, Chief Technology and Data Officer at Allianz Spain, put it bluntly: "AI has both a light side and a dark side. While most focus on the opportunities, it also introduces new vulnerabilities, and many organizations are more exposed than they realize."

IBM's analysis shows that in organizations relying on manual governance, incident risk increases as AI adoption scales. Conversely, those that embed control directly into their AI systems experience 25% fewer incidents.

Most surveyed tech CxOs (59%) cite security and compliance concerns as top barriers to scaling AI agents—not technical capability, but the inability to manage risk at scale.

The Financial Blind Spot: $1 in Every $4 by 2027

Here's the CFO nightmare: AI spend is becoming a material budget line—25% of IT budgets by 2027—while 84% of technology leaders have not operationalized AI financial management.

This isn't about forecasting error. This is about fundamental visibility. 85% of surveyed executives lack real-time AI spend tracking. They don't know what models are running, what inference costs are accruing, or which departments are consuming capacity until the bill arrives.

The IBM study found that organizations with strong financial discipline deploy 2.4x more AI agents with no higher AI/IT budget. They're also 3x more likely to say they are fully prepared for AI scale. Financial control isn't slowing them down—it's enabling them to scale faster.

The penalty for flying blind? Organizations that build control into their AI systems:

• Deploy 16x more AI agents than those relying on manual governance
• Deliver 18% higher operating margins
• Spend 4x less of their AI budget on wasted capacity or rework

Those are board-level performance differences, driven entirely by structural discipline.

The Adaptability Premium: Portable Workloads, Replaceable Models

Beyond governance and financial control, the study reveals a third dimension: adaptability. Organizations that designed for vendor flexibility early—keeping workloads portable and models replaceable rather than locked into hard dependencies—reported a 10% higher return on AI investment in 2025.

Boris Alexandre, Head of ARP Programme at Airbus, described their approach: "We design modular architectures so components can evolve as technology advances, without breaking the overall system. That approach allows us to absorb rapid innovation while supporting products with decades-long lifecycles."

This isn't theoretical. With model pricing, performance, and licensing terms shifting quarterly, organizations that hard-coded dependencies on specific vendors are now re-architecting under pressure. Those who built for portability from day one are swapping in better models without disruption.

What CIOs, CTOs, and CFOs Should Do Now

The IBM study makes clear that control, visibility, and adaptability are not brakes on AI deployment—they are accelerators. Organizations that build these capabilities early deploy more agents, achieve higher margins, and prepare for scale with confidence.

For CTOs: Embed Control into Architecture

Stop governing AI manually. Manual approval workflows, spreadsheet tracking, and quarterly reviews cannot keep pace with autonomous systems. Instead:

1. Build observability into your AI systems from day one. Instrument model calls, track inference costs in real-time, and flag anomalies automatically.
2. Deploy guardrails at the architecture level, not the approval layer. If an agent shouldn't access customer PII, enforce that in code, not policy.
3. Design for portability. Use abstraction layers that allow you to swap models, vendors, or inference platforms without rewriting applications.

Organizations that embed control at the architecture level deploy 16x more agents than those relying on manual governance—because they don't need to review every deployment individually.

For CIOs: Rebuild Governance for Continuous Deployment

Your governance model was designed for predictable release cycles. AI agents operate continuously, autonomously, and across organizational boundaries. That requires a different governance approach:

1. Shift from approval-based to observation-based governance. You can't approve every model call or agent interaction. Instead, define acceptable behavior, monitor deviations, and intervene when systems drift.
2. Create a control plane for AI—a central visibility layer that tracks what's deployed, what it's costing, and what business outcomes it's delivering, regardless of which team or department initiated it.
3. Focus on incident response, not incident prevention. With 54 AI agent incidents per year on average, the question isn't "how do we prevent all incidents?" but "how quickly can we detect and contain them?"

Organizations with strong governance maturity experience 25% fewer high-severity incidents and are 3x more likely to say they're ready for AI scale.

For CFOs: Treat AI Like Labor, Not Software

AI spend is becoming a material budget line—25% of IT budgets by 2027—but most finance teams still manage it like a software license. That doesn't work when spend is usage-driven, distributed across departments, and growing 71% in two years.

1. Implement real-time AI cost tracking. If you can't see spend until the bill arrives, you can't manage it. Instrument model calls, track inference costs by department, and flag runaway consumption before it hits P&L.
2. Benchmark AI spend against labor economics. If an AI agent costs $12,000/year to run but replaces $80,000 in contractor spend, that's a 6.7x ROI. If it costs $12,000 and saves 2 hours/week for a $60,000 employee, that's breakeven at best.
3. Model AI budgets as portfolio investments, not individual projects. Organizations that manage AI as a portfolio deploy 2.4x more agents with no higher AI/IT budget, because they optimize across use cases rather than optimizing each one in isolation.

The IBM data is unambiguous: organizations with strong financial discipline are 3x more likely to say they're fully prepared for AI scale. Financial control isn't overhead—it's the foundation for scaling with confidence.

The 2026 Accountability Shift: From IT to the C-Suite

The most significant finding in the IBM study may be the accountability shift. 80% of respondents report CEO-driven AI transformation mandates—not CTO-driven, not CIO-driven. AI is no longer a technology initiative delegated to IT. It's a board-level transformation that CEOs are driving directly.

That changes the stakes. When the CEO owns AI transformation, 67% of CIOs and CTOs being accountable for systems they don't fully control is no longer a technical problem. It's a governance crisis that threatens execution on a CEO mandate.

The organizations winning in this environment aren't the ones deploying the most agents. They're the ones deploying agents with control, visibility, and adaptability built in from day one. They're the ones that can answer the CEO's question: "How much are we spending on AI, what are we getting for it, and can we scale this 10x without breaking the business?"

Matt Lyteson, CIO at IBM, summarized the challenge: "It is no longer just about deploying AI faster. It's redesigning how organizations control, govern and invest in it—and embedding control and visibility from the start, so they can scale with confidence."

The control gap isn't closing by accident. It's closing because technology leaders are redesigning their architectures, governance models, and financial processes to handle systems that operate continuously, autonomously, and at enterprise scale.

The question for CIOs, CTOs, and CFOs isn't whether to build these capabilities. It's whether to build them now—while you're scaling from 50 agents to 500—or later, when you're trying to govern 5,000 agents with spreadsheets and manual approvals.

Sources

1. New IBM Study Finds CIOs and CTOs Face Growing AI Control Gap as Enterprise Deployment Scales - IBM Newsroom, June 8, 2026
2. 2026 Tech Leader Study: Building the IT foundation for agentic AI at scale - IBM Institute for Business Value, 2026