Why most IT teams still struggle to see real ROI from AI

A significant portion of AI initiatives in IT infrastructure & operations fail

AI has great potential, but most enterprise projects still miss the mark. Gartner’s survey of 783 infrastructure and operations leaders shows that only 28% of AI projects meet ROI expectations. Another 20% fail completely. These numbers highlight a persistent gap between ambition and execution. Too often, teams run AI experiments without a clear business goal, no alignment with strategy, and limited technical readiness.

For executives, this means one thing, AI must be treated as a core strategic investment. Success depends on clarity of purpose and integration into the business’s operating structure. When AI is deployed only at the edge of operations, its benefits stay limited to localized gains. But when the organization commits to embedding it into core systems and processes, returns multiply.

Leaders must also face a skills reality. Many organizations lack the expertise to train, maintain, and scale AI systems effectively. Without strong engineering and data governance, AI initiatives create cost without outcome. Building internal capability or securing external expertise is is fundamental to success.

According to Gartner’s findings, 57% of respondents had at least one AI failure. That figure doesn’t mean AI isn’t worth it. It means most organizations are still learning how to manage it. Approach it with discipline, measurable targets, and accountability. AI that starts with a defined business purpose and is integrated into daily operations will outpace speculative projects every time.

Unrealistic expectations and poor integration practices are major causes of AI project failures

Many leaders still expect that AI will instantly automate complex tasks and deliver cost savings without much effort. It doesn’t work that way. Melanie Freeze, Director of Research at Gartner, said organizations often expect “too much, too fast.” When early AI models don’t perform miracles, confidence drops and projects stall.

AI needs time to generate results. It must be integrated gradually into processes that employees already use and understand. Poor integration often leads to limited adoption and abandoned tools. When technology doesn’t fit the workflow, users fall back on tradition, and ROI disappears.

For executives, this is a leadership challenge more than a technical one. You need to calibrate expectations. Define success metrics that are realistic, measurable automation gains, operational accuracy, and scalability over months, not days. Make sure teams understand the goal isn’t to replace people overnight, but to amplify how your systems and staff perform.

What truly separates successful AI strategies from failures isn’t the sophistication of the model, it’s alignment with the business. When AI decisions are made close to operations, with clear governance and resource backing, ROI follows. When expectations are driven by hype, investments vanish into pilot projects that never scale.

C‑suite executives should focus on integrating AI into the company’s strategic plans and track progress as they would any other product or service initiative. AI returns improve when leadership manages momentum with clear communication, patient execution, and defined performance metrics.

Integrating AI into existing business processes with full executive sponsorship is critical for achieving ROI

The most successful organizations embed AI into systems that people already rely on. It’s not about launching isolated projects; it’s about extending core operations with intelligent automation and decision support. Gartner’s research highlights three major success factors, make AI part of day-to-day processes, secure visible executive backing, and create realistic, outcome-based business cases.

Melanie Freeze, Director of Research at Gartner, emphasized that AI use cases must be managed as products. This means tracking cost, scalability, and measurable business impact. AI initiatives that are built and governed this way generate consistent value because they remain aligned with existing operational and financial priorities.

Executives play a vital role in this process. When leadership actively supports AI adoption, it removes the usual roadblocks, funding uncertainty, fragmented priorities, and internal resistance. A clear chain of ownership also ensures that new technology isn’t treated as an optional experiment but as part of the company’s long-term operational fabric.

A shared scoring model, as Freeze suggests, helps leaders compare projects across departments. It allows them to decide which cases deserve higher funding and which should be delayed or redesigned. This disciplined approach ensures return on AI investment isn’t an accident, it’s planned, measured, and continuously optimized.

Executives who want AI to deliver must make its success structural. It should have a defined road map, committed sponsorship, and visible integration into the business engine.

The most tangible AI successes are found in mature areas such as IT service management (ITSM) and cloud operations

AI doesn’t need to start from scratch to deliver value. Gartner’s findings show that 53% of I&O leaders report successful results when applying AI within IT service management (ITSM). These areas have structured data, well-defined workflows, and measurable service goals, ideal conditions for AI to demonstrate clear ROI.

Melanie Freeze points out that I&O leaders find the strongest results where markets are already stable and operational frameworks are established. ITSM and cloud operations provide a controlled environment for AI to optimize performance and reliability. Success in these fields is easier to scale because teams understand the system behavior, data flow, and performance benchmarks from the start.

For executives, the message is simple: start where the path is proven. Focus on practical, high-value use cases with measurable outcomes. As early successes generate confidence and institutional learning, scaling AI into new areas becomes faster and more predictable.

Mature domains also help organizations refine governance and compliance standards before expanding to more complex applications. When AI runs in well-managed environments like ITSM, results are transparent and repeatable, key attributes executives need to justify further investment.

This strategy doesn’t limit innovation; it builds a stronger foundation for it. Success in ITSM and cloud operations gives organizations the operational discipline and data maturity needed to apply AI more broadly across the enterprise.

A cohesive, organization‑wide strategy with centralized funding oversight is essential to mitigate risks and optimize AI investments

AI investments fail most often when each business unit runs its own isolated experiments. Without coordination, projects overlap, compete for resources, and operate with misaligned objectives. Gartner’s analysis shows that successful organizations treat AI as a unified strategic portfolio, driven by a common framework for funding, evaluation, and governance.

A centralized approach ensures fewer redundant efforts and stronger scalability. It allows executives to allocate resources where they have the greatest impact and manage risk across departments. This governance model also enables clearer accountability, leaders can track the contribution of specific AI projects to operational outcomes.

Executives should view AI funding as a long‑term capital strategy. Gartner emphasizes that CEOs and CFOs need to play a more active role in defining funding criteria and approving major AI investments. Their oversight helps ensure that spending aligns with business priorities, especially as infrastructure and data costs increase.

This structure also supports transparency. With a unified scoring system to assess feasibility, cost, and expected impact, leadership can make better decisions about where to invest. It eliminates guesswork and ties each initiative to measurable business and financial value.

Organizations that centralize AI governance gain better control over scale, consistency, and performance. Executives who enforce this model can focus company resources on the projects that drive growth, resilience, and long‑term returns.

Failed AI initiatives can harm organizational performance and credibility by undermining infrastructure reliability

When AI projects fail, the cost goes beyond financial loss. Gartner’s findings indicate that unsuccessful deployments can damage IT reliability, security, and service availability. These issues disrupt operations, slow innovation, and weaken stakeholder trust. For any enterprise, maintaining stability while adopting AI is essential for reputation and long‑term competitiveness.

Melanie Freeze, Director of Research at Gartner, stressed that the impact of failed AI isn’t just technical, it’s organizational. Poorly executed AI projects consume resources and reduce confidence across teams, making it harder to justify future innovations. To avoid this, she recommends a clear strategic foundation supported by robust governance and execution discipline.

Executives should ensure every AI initiative begins with a defined business case. It must address specific operational problems and have a measurable performance goal. Projects that lack this alignment create more risk than reward. Proper planning, realistic expectations, and continuous evaluation are what sustain AI credibility and results.

For leadership, this is not just about protecting investments, it’s about maintaining organizational integrity. A failed deployment can compromise system availability or data compliance, affecting enterprise continuity. Avoiding these pitfalls requires central oversight and a resilient adoption framework that keeps reliability as a top priority.

Strong execution, business adoption, and consistent follow‑through are the foundation of meaningful ROI. AI success is not achieved by prioritization alone, it depends on disciplined implementation and continuous accountability.

Key takeaways for decision-makers

• AI ROI remains limited for most IT organizations: Only 28% of AI initiatives in infrastructure and operations meet ROI expectations. Leaders should focus on integrating AI strategically across business functions instead of running isolated experiments.
• Unrealistic expectations drive many failures: Many IT teams expect immediate automation and cost savings. Executives should set realistic goals, communicate clear performance metrics, and build in time for gradual adoption.
• Embedding AI and securing leadership support boosts success: Align AI with existing workflows and make executive backing visible. Leaders should treat AI initiatives as managed products with clear business cases and accountable ownership.
• Start with mature, high‑value domains like ITSM: Most successful AI implementations, 53% according to Gartner, occur in IT service management and cloud operations. Decision‑makers should prioritize mature areas before scaling enterprise‑wide.
• Centralized AI strategy and funding strengthen outcomes: Decentralized efforts waste resources and dilute impact. CEOs and CFOs should enforce centralized governance and unified scoring models to fund projects with the highest strategic value.
• Poorly executed AI harms infrastructure and credibility: Failures can erode trust, disrupt operations, and weaken IT reliability. Executives must demand disciplined execution, clear ROI metrics, and a robust business case for every AI deployment.