A new IBM study delivers a sobering reality check for enterprise AI: 66% of CIOs and CTOs report being held accountable for AI systems they do not fully control. As AI moves from pilot programs to enterprise-wide deployment, the gap between responsibility and visibility is widening—and the consequences are landing squarely on technology leaders' desks.
The IBM Institute for Business Value study surveyed 2,000 C-level technology executives across 33 countries and 19 industries from January to April 2026. The findings paint a picture of AI adoption that's outpacing the organizational structures designed to support it.
The Control Gap: What's Actually Happening
Here's the problem in numbers:
- 70% of surveyed executives say teams across the business are deploying technology faster than IT can track
- 77% report AI adoption is already outpacing current governance capabilities
- Only 11% believe they are fully ready for the scale of AI agent deployment expected in the next year
- 84% have not fully operationalized AI financial management
- 85% still lack full visibility into real-time AI spend
The math is brutal. By 2027, surveyed technology leaders anticipate a 38% increase in the number of AI agents deployed. Meanwhile, 80% report CEO-driven AI transformation mandates pushing them to scale faster—even as the structures to support that scale lag behind.
As Matt Lyteson, CIO at IBM, puts it: "For CIOs and CTOs, the challenge now is scaling AI systems that operate continuously and autonomously, often within governance models and architectures designed for a far slower, more predictable environment."
The Real Cost of the Control Gap
This isn't just a governance theory problem. The control gap has real operational and financial consequences.
Incident Risk
Organizations surveyed experienced an average of 54 AI agent incidents last year—situations where an unintended or harmful occurrence required human correction. Of those:
- 17% were high severity, requiring more than four hours to contain
- 37% resulted in data exposure or security breaches
- 33% caused cascading system failures
- 17% triggered compliance issues
The IBM analysis shows that in organizations relying on manual governance, incident risk increases as AI adoption scales. In contrast, organizations that embed control directly into their AI systems experience 25% fewer incidents.
The Financial Blind Spot
AI spending is projected to grow from just under 15% of IT budgets in 2025 to nearly 25% by 2027—a 71% increase in two years. For large enterprises, this translates to an average of $186 million annually on AI, according to recent industry data.
Yet most organizations can't answer the fundamental CFO question: What did we get back?
- Only 8% of enterprises have achieved meaningful business returns with AI (KPMG)
- 74% of organizations want AI to grow revenue, but only 20% have seen it happen (Deloitte)
- Only 27% of executives say AI has met their ROI expectations
- Token spend is up 13x since January 2025, yet ROI visibility remains opaque
"Companies aren't failing because AI doesn't work," says Alina Vrsaljko, CEO of Botanu, a startup focused on AI performance measurement. "They're failing because they can't locate where their agents are working. 72% of CEOs now own the AI decision, and they have no way to prove it's paying off."
Why Traditional Governance Breaks at AI Scale
The cloud computing playbook doesn't translate to AI. In the cloud era, cost increased with usage, and every bill could be mapped back to the workload, team, or product that drove it. AI breaks that model.
The same task can produce very different costs from one run to the next, with little predictability upfront. Pricing has shifted from flat per-seat subscriptions to usage-based models, pushing volatility onto the buyer's invoice. By the time the bill arrives, no one can tell which agents were actually worth it.
AI agents operate across multiple systems—each metered differently, each owned by a different team. A single agent's cost is scattered across model vendors, tool APIs, and infrastructure layers. Its value is equally distributed: Did the sales agent lift revenue? Did the customer service agent solve tickets successfully? Did it protect EBITDA?
Traditional governance treats AI like software licenses. But as Vrsaljko frames it: "An AI agent is a new kind of workforce, and it works at 100 times the frequency of a person. You should performance-manage it, not just cost-manage it."
The Organizations That Are Getting It Right
The IBM study also identifies what separates high-performing AI organizations from the rest. The differentiator isn't spending more—it's designing control into AI systems from the start.
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 (per unit of value created)
Organizations with strong financial discipline:
- Deploy 2.4x more AI agents with no higher AI/IT budget
- Are 3x more likely to say they are fully prepared for AI scale
Organizations that designed for adaptability early—keeping workloads portable and models replaceable rather than locked into hard dependencies—reported a 10% higher return on AI investment in 2025.
The pattern is clear: control scales, manual governance doesn't.
What "Outcome-Maxxing" Means in Practice
Industry observers are calling for a shift from "token-maxxing" (measuring activity) to "outcome-maxxing" (measuring results).
"Activity is not outcome," says Deborah Jacob, CTO at Botanu. "A thousand tokens and ten tool calls tell you an agent was busy—not whether it closed the deal. We measure the result the business actually recorded, and weigh it against what it cost to get there—the one number a CFO can act on."
The companies doing this right get approximately $3 back for every $1 invested in AI. Right now, only about 6% of companies lead with that discipline.
Ray Rike, CEO of Benchmarkit and host of the "AI to ROI" podcast, frames it this way: "The discipline that's missing is simple to say and hard to do: measure outcomes, not activity, and connect costs to returns. Until then, AI is a compensation-scale expense that demands CFO-level governance most companies don't yet have."
What CIOs and CTOs Should Do Now
The IBM study offers three structural recommendations for technology leaders:
1. Redesign Control, Don't Retrofit It
Embed visibility, governance, and financial tracking into AI systems at deployment—not as an afterthought. Organizations that design control in from the start experience 25% fewer incidents and deploy 16x more agents than those relying on manual oversight.
Tactical steps:
- Instrument AI agents with telemetry that captures full activity across model vendors, tools, and infrastructure
- Build real-time dashboards that show AI spend by agent, team, and business outcome
- Create agent-level accountability: each agent should have an owner, a job description, and success metrics
2. Operationalize AI Financial Management
Treat AI spend like labor cost, not software licensing. Every agent should have a cost-to-outcome ratio that a CFO can evaluate.
Tactical steps:
- Map AI spending to business outcomes recorded in systems like CRM, support ticketing, or financial platforms
- Compare agent performance to the cost of human labor for the same task
- Establish ROI thresholds: kill agents that don't meet them, scale agents that exceed them
3. Design for Adaptability
Keep workloads portable and models replaceable. Hard dependencies lock you into vendors and prevent optimization as the landscape shifts.
Tactical steps:
- Use modular architectures that allow components to evolve without breaking the overall system
- Avoid vendor lock-in by maintaining abstraction layers between business logic and model APIs
- Regularly benchmark alternative models and providers to validate you're getting competitive value
As Dalton Gouws, Group IT Director at VWG UK, puts it: "We don't know who's going to win or lose over the next five years. So we're keeping AI models plug-and-play, ready to adapt if the landscape shifts."
The Boardroom Question That's Coming
This fall, when budget planning hits, a lot of leaders will face the same question: What did we get for $186 million in AI spend?
The ones who can answer that question—with data, with business outcomes, with a credible ROI story—will pull ahead. The ones who can't will face budget cuts, executive skepticism, and pressure to prove value before scaling further.
The control gap isn't a technical problem. It's a structural one. And it's solvable—but only if technology leaders redesign how organizations govern, measure, and invest in AI before the gap becomes a crisis.
As Afonso Eça, Executive Board Member at Banco BPI, describes 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. No one would choose to pilot that plane—but that's exactly what companies are doing today."
The question for CIOs and CTOs isn't whether to scale AI. The mandate is already here. The question is whether they'll have the visibility, control, and financial discipline to do it without crashing.
Sources:
