Information technology executives aiming to optimize costs and boost efficiency using AI in IT infrastructure should be aware: just 28% of initiatives achieve full return on investment. A Gartner survey highlights that 20% of AI projects in IT infrastructure and operations don’t succeed, while 57% of IT managers experience at least one failure.

Melanie Freeze, Gartner’s research director, suggests these failures often stem from unrealistic assumptions. Many anticipate immediate automation of complex tasks, significant cost reductions, or solutions to long-standing issues, which leads to disappointment and halts progress.

Issues such as poor planning or overly ambitious goals contribute to the 20% failure rate. AI solutions that don’t integrate seamlessly into existing operations fall short on delivering ROI.

Failures are notably frequent in auto-remediation, self-healing infrastructure, and agent-driven workflow management, as per Gartner findings. Among IT executives who’ve faced obstacles, 38% acknowledge ongoing skill shortages, while another 38% cite inadequate data quality or availability.

However, the picture brightens in areas where AI technology is mature, such as IT service management (ITSM) and cloud operations, where 53% of leaders report success.

Funding for AI in tech infrastructure remains a challenge, often reliant on individual business units’ budgets. As expenditures grow, Gartner urges top executives to engage in investment decisions related to AI funding.

Contextualizing these findings is crucial against the larger backdrop: a February survey of nearly 6,000 executives across multiple countries found over 80% saw no tangible impact from AI on employment or productivity, though 69% of businesses utilize AI.

Pressure to show AI investment returns is on the rise, highlighted by a Harris Poll survey indicating that 98% of tech leaders face scrutiny from boards to prove ROI, with 71% of CIOs anticipating budget cuts if targets aren’t met by mid-2026.