AI often doesn’t deliver ROI for IT departments either

Results of  a Gartner study released Tuesday reveal that only 28% of AI use cases in infrastructure and operations (I&O) fully succeed and meet ROI expectations, and a full 20% end up failing outright.

According to Melanie Freeze, a director of research at Gartner, failure “most commonly occurs” for several reasons, including unrealistic expectations of what AI tools can do, and skills gaps during the actual pilot.

While these results are an improvement over the troubling findings from MIT released last year that revealed 95% of genAI projects produce no measurable financial return, there is, she said in an interview with CIO.com, a great deal of experimentation going on among IT departments in which a team of I&O professionals will “just go out and try something.”

The reality, said Freeze, is that in order to achieve an anticipated ROI, IT departments must not opt to run them as side projects.

In a Gartner Q&A advisory about the survey of 783 I&O leaders conducted late last year, she stated that, of the 57% of I&O leaders reporting at least one failure, “many said their AI initiatives failed because they expected too much, too fast. They assumed AI would immediately automate complex tasks, cut costs, or fix long‑standing operational issues. When expectations are not realistically set and the results don’t appear quickly, confidence drops and projects stall.”

The survey, she said, revealed that ROI from AI is not driven by the sophistication of the model, but by how well the technology is integrated, governed, and aligned with real operational needs.

Success factors

To that end, Gartner said it has identified what it calls three success factors. These include embedding AI into the systems and processes people already use. “As AI becomes part of day‑to‑day operations, it boosts adoption and creates visible impact within the organization,” the company noted.

Successful I&O executives also receive full support from top executives, which helps “remove roadblocks, align priorities, and ensure the investment stays funded and focused,” and they create realistic business cases.

Freeze said that I&O leaders should prioritize and determine funding for AI use cases “by managing AI use cases as a product to avoid duplication, drive synergies, and track their collective impact on I&O and business outcomes.

From there,” she said, “I&O leaders can work alongside their CIOs, data and analytics, security, legal, and finance stakeholders to assess each use case for feasibility, risk, cost, and expected business impact. A shared scoring model makes it easy to compare and rank all use cases and guide investment decisions.”

She pointed out that the bulk of the success comes from genAI applied to specific areas: IT service management (ITSM) and cloud operations, “where markets are mature and have proven business value. In fact, 53% of I&O leaders reported their AI wins occur in ITSM,” she noted. “Whether these wins occur in the cloud or in ITSM, I&O leaders must ensure they are shared broadly within the organization, and the AI strategy remains cohesive and centrally led.”

Needs to be grounded in a business case

Starting without a plan, she told CIO.com, is never a good idea: “It’s always a bad situation for any technology to say, ‘we built it. It’s going to succeed.’ It needs to be grounded in the business case. What does your business need? What are their ambitions? What are the problems within your function that your current tool set is not able to solve? Within that upfront strategic framework, then success follows.”

There is also the problem that a failed AI project can affect an entire organization. Not being able to provide secure, reliable, available infrastructure can have major implications for business outcomes, said Freeze.

“The drivers of failure are slightly different from the drivers of success,” she said. “I&O leaders must remember that a clearly defined, centrally endorsed AI portfolio helps their organization focus resources where they matter most. Above all, strong execution and business adoptions, not just prioritization, determine AI’s real ROI.”

Once priorities are clear, added Freeze, they can then determine which use cases deserve funding and at what level. “Today, many AI initiatives are still funded by individual business units,” she observed. “However, as AI infrastructure spending continues to rise, CEOs and CFOs need to play a more active role in setting funding criteria and approving major investments.”