Every CIO I talk to is under the same pressure: prove AI is working before the board loses patience. The math still isn't adding up for most of them.

Global AI spending is projected to hit $2.59 trillion in 2026 — a 47% jump from the prior year. Boards are demanding readouts. Budgets are being approved. Tools are being deployed.

Yet according to CIO.com's 25th annual State of the CIO survey — 662 IT leaders and 249 line-of-business users — only 19% say AI initiatives have met or exceeded business goals. Eighteen percent admit fewer than a third of their AI use cases are meeting defined expectations.

Eighty-one percent of CIOs are missing their targets. The technology isn't the problem. The foundation is.

The Five Blockers Stalling AI ROI

Three primary barriers dominate the State of the CIO 2026 data. Understanding them is the first step toward escaping them.

Lack of in-house expertise (40%) is the single biggest obstacle. Organizations can buy tools faster than they can build teams. Data scientists, AI engineers, prompt specialists, and governance experts take months to hire and years to develop. Without them, enterprise AI deployments rely on vendor-provided implementations that can't adapt to organizational context.

Ill-defined ROI metrics (32%) create a measurement trap. "No one is measuring ROI on an ongoing basis because we are facing counterpressures from every VP and line-of-business domain looking to implement AI for their own optimization," notes the CIO at Rensselaer Polytechnic Institute. "We are saying yes to everyone without stepping back and focusing on the business cases that show real value."

When every department has an AI project and no one owns the portfolio-level business case, you end up with dozens of pilots that each look promising in isolation but collectively fail to move the needle on enterprise performance.

Murky corporate AI strategy (31%) is the third blocker. AI projects multiply while enterprise strategy lags. Individual departments move fast, but without a coherent AI roadmap, organizations accumulate redundant tools, fragmented data pipelines, and competing governance standards that slow everything down.

Two additional blockers sit below the surface but may be even more structural:

Data quality and infrastructure gaps surface problems that already existed in your data layer. Inconsistent schemas, siloed data lakes, legacy systems without API access — these don't just slow AI deployment. For some use cases, they make deployment impossible until the underlying data architecture is modernized.

Security and governance immaturity becomes critical with agentic AI. An autonomous agent that can browse the web, write code, and call external APIs needs fundamentally stronger controls than any prior enterprise software category. Most organizations haven't built those controls yet.

Why CIOs Are Returning to Fundamentals

Info-Tech Research Group's Best of 2026 Mid-Year Report — which tracks the most-accessed research by IT leaders globally — reveals something counterintuitive: the most popular resources aren't about deploying the latest models. They're about going back to basics.

The top topics driving demand among IT leaders in the first half of 2026:

• AI execution and operationalization
• Data strategy and information management
• Cybersecurity resilience
• Infrastructure modernization
• Enterprise risk management
• Workforce development and readiness
• Vendor evaluation and technology buying decisions

"AI is no longer at the edge of the CIO agenda as an experiment. It is becoming part of how organizations plan, operate, secure, and deliver value," says Gord Harrison, Chief Research Officer at Info-Tech Research Group. "The most accessed resources from the first half of 2026 show that IT leaders understand the work ahead. They are strengthening data, governance, security, infrastructure, risk, and workforce capabilities so AI can scale responsibly instead of adding complexity."

This is the quiet shift happening beneath the AI headlines. While vendors pitch agents and foundation models, CIOs are spending their calendar hours on data governance frameworks, zero-trust security for agentic systems, infrastructure capacity planning, and workforce readiness programs.

Not because they're AI skeptics. Because they've been burned by deploying too fast on a foundation that wasn't ready — and they're not making the same mistake twice.

The Info-Tech research maps directly onto the five AI priorities that determine whether AI scales or stalls. Their most-demanded resources span the same territory: AI implementation guidance, information management, cybersecurity, infrastructure, enterprise risk, workforce development, and vendor evaluation. These aren't peripheral IT concerns. They're the determinants of AI success at scale.

The Gartner Finding That Reframes the ROI Conversation

Gartner's analysis of 101 "efficient growth leaders" versus matched control peers — presented at the Gartner Finance Symposium/Xpo 2026 — adds a critical strategic dimension that most CIOs are missing: it's not how much you spend on AI. It's where you deploy it.

The data is striking:

• 46% of efficient growth companies deploy AI across both product innovation AND sales, marketing, and customer growth — versus 32% for control group peers
• Companies below $3 billion in revenue: efficient growth leaders deployed 2x more AI use cases than comparable peers at the same revenue scale
• Companies below $10 billion in revenue: efficient growth leaders were 2.6x more likely to deploy AI across both product and customer engines simultaneously
• AI competitive advantage was strongest in data-intensive technology and financial services, where AI embeds directly into products and customer interactions

"Simply spending more on AI does not, by itself, equate to better business outcomes," said Michelle Carlsen, Director Analyst in Gartner's Finance practice. "Organizations that outperformed industry peers on revenue growth, margin expansion, and return on invested capital over the last ten years were more likely to frame AI as a growth engine and to connect AI use cases across product innovation and sales, marketing, and customer growth."

The most important finding in the Gartner research: "Productivity-focused AI investments alone do not explain performance differences, and automation by itself is increasingly becoming table stakes rather than a durable source of advantage."

This reframes the entire CIO mandate. If your AI strategy is primarily focused on cost reduction and internal productivity — which is where most enterprises start — you're executing a commodity play. The leaders are integrating AI into product development and customer growth simultaneously, creating compounding advantages that efficiency-only organizations cannot replicate on any timeline.

What the Top 19% Do Differently

The State of the CIO 2026 data reveals structural patterns among the minority of CIOs whose AI initiatives are hitting their marks.

Cross-functional governance with real accountability. Eighty-three percent of IT leaders surveyed have or plan to establish cross-functional AI steering committees. But the high performers aren't just creating committees — they're designing governance that forces accountability at the point of business impact.

First Student, a school transportation provider, operates a well-defined AI council with C-suite and business leader representation that meets regularly to review use cases against financial business cases. "We have more discipline around business cases than most companies," says CIO Sean McCormack. "Everything is metrics-driven and dependent on proving value. By the time we put something into production, it's been through a series of proof of concepts, there's been a deep dive on financials, and we are able to move quickly and demonstrate value."

Replacing centralized AI Centers of Excellence with embedded squads. The CoE model — favored by enterprise architecture teams — is showing its limits in 2026. Centralized CoEs create a clearinghouse that nobody ultimately owns. High-performing organizations are replacing them with AI squads embedded inside individual business units. Embedded teams force accountability at the point of business impact rather than in a central function one degree removed from actual outcomes.

Joint technical and business ownership on every project. Each AI initiative gets a named technical sponsor and a named business sponsor, both co-owning outcomes. This eliminates the classic failure pattern where IT builds something the business doesn't fully use, and neither party is accountable for the gap.

KPIs established before deployment, not after. Only 47% of organizations have established formal AI success metrics — with another 34% planning to within the year. The top performers are in the 47% who have already done this. They're measuring what matters: operational efficiency and process improvement (40% prioritize this), employee productivity (34%), and direct cost reduction (30%). Revenue impact is a less immediate priority (27%), suggesting the 2026 AI maturity curve is still weighted toward internal efficiency as the first validation gate.

TIAA's honest assessment matters here. Despite having 85% of its workforce using an internal AI platform, robust governance frameworks, steering committees, an AI CoE, and an enterprise mandate for strategic use of AI embedded in every employee's performance goals, their chief operating, information, and digital officer acknowledges: "What's on paper sometimes doesn't turn into real ROI given the reality of operational costs. Something may prove to be a successful pilot, but you need to understand the full cost of operations — for example, the efficiencies of running tokens or how you're handling traffic."

Even with all the right structures, execution details matter. Token costs, infrastructure overhead, and operational maintenance costs can erode pilot-stage ROI projections by 30-50% in production environments.

The CFO Lens: Strategy Over Budget Size

For CFOs evaluating AI investment portfolios, the Gartner research delivers a clear framework that should reshape how AI budgets are structured and measured.

Don't evaluate AI investments primarily by the return of individual use cases. Evaluate them by how well those capabilities reinforce broader growth, product, and decision processes across the enterprise.

Three practical implications for CFO decision-making:

AI productivity investments alone won't differentiate your business. Your competitors are buying the same productivity tools, deploying the same code assistants, and automating the same back-office workflows. Differentiation comes from AI embedded in products and customer relationships — territory that most enterprises haven't entered yet.

Company scale no longer equals AI advantage. The Gartner data is clear: companies below $3 billion in revenue are deploying AI as a scale multiplier with outsized results versus larger peers. The era of "AI is only for large enterprises with massive data teams" has ended. Mid-market companies that build AI-native products now have a credible path to outperform larger competitors who are stuck automating legacy processes.

Industry context determines your current AI leverage point. In asset-intensive industries — manufacturing, logistics, utilities — AI is primarily delivering efficiency gains right now, not differentiation. The competitive pressure is real but the advantage is modest. In data-intensive industries — financial services, software, media — AI is already creating durable competitive moats. Your AI prioritization framework should reflect which type of leverage is available to you today.

The CTO Build Mandate: Foundations Determine Scale

For CTOs, the Info-Tech demand data maps directly onto five foundation capabilities that determine whether AI scales or permanently stalls at the pilot stage.

Data practices come first. AI output quality is bounded by data input quality. Organizations with clean, well-governed, accessible data can move to production AI in weeks. Organizations with fragmented, inconsistently formatted, siloed data spend months cleaning before they can deploy — and often discover the data problem is larger than the AI project itself.

Infrastructure must be purpose-built for AI workloads. Legacy infrastructure wasn't designed for token-based billing, GPU compute demands, or the network requirements of agentic AI systems. CTOs who treat AI infrastructure as a software problem (model selection, API integration) while leaving hardware capacity and cost management to chance are setting up for budget surprises at scale.

Risk management must be integrated, not bolted on. AI risk is simultaneously IT risk, business risk, regulatory risk, and reputational risk. CTOs who treat AI governance as a security checklist will consistently miss the broader risk surface. Effective AI risk management requires cross-functional ownership — security, legal, compliance, and business unit leaders all in the loop before deployment, not after an incident.

Zero-trust security for agentic systems is non-negotiable. Agentic AI systems that can browse the web, write and execute code, and call external APIs introduce attack surfaces that most enterprise security frameworks weren't designed to address. The traditional perimeter model doesn't cover AI agents acting autonomously on behalf of the organization.

Adaptive leadership is an operational requirement. CTOs must manage two speeds simultaneously: governing and optimizing today's AI implementations while building the capabilities for tomorrow's AI-native architecture. Organizations that let current operational demands crowd out foundational capability-building will be in the same position in 2027 that they're in today — always ready to scale AI, never quite getting there.

The Bottom Line

Eighty-one percent of CIOs are missing AI ROI targets in 2026. The technology isn't the constraint. The failure is about sequence — deploying AI before the foundation that makes it scale.

The top 19% share a common pattern: strong data governance built before deployment, accountability embedded at the business unit level rather than centralized in a CoE, and AI deployed strategically across both efficiency (necessary but commoditizing) and growth (where differentiation actually lives).

Gartner's efficient growth leaders aren't spending more than their peers. They're deploying more strategically — integrating AI across product innovation and customer growth simultaneously, creating mutually reinforcing capabilities that productivity-focused competitors cannot replicate.

If your AI strategy today is primarily about automating internal processes and reducing headcount, you're building the floor, not the ceiling. The companies winning with AI in 2026 are treating automation as table stakes and racing toward product and customer differentiation as the real prize.

Sources

• State of the CIO 2026: CIOs Set the Course for AI ROI — CIO.com / Foundry, 662 IT leaders + 249 LoB respondents
• AI Execution Is Pushing CIOs Back to IT Fundamentals — Info-Tech Research Group, Best of 2026 Mid-Year Report
• Gartner Says CFOs Gain a Competitive Advantage from Strategic AI Deployment, Not AI Spending Levels — Gartner Finance Symposium/Xpo 2026, analysis of 101 efficient growth leaders