Only 13% of chief information officers (CIOs) said their organisations had achieved significant or transformational value from artificial intelligence, even as companies ramp up spending and expand its use, according to a Gartner survey.
The survey of 2,002 CIOs and technology executives, conducted from May to July 2026, also found that only about one in five AI use cases succeeds.
The findings point to a growing challenge for companies investing in AI: turning experiments and deployments into measurable business results.
Alicia Mullery, a vice president analyst at Gartner, said companies needed to establish clear targets for AI projects before committing resources and be prepared to abandon initiatives that fail to deliver.
Organisations with guidelines for deciding when an AI pilot should not proceed to production were 2.3 times more likely to be among Gartner’s top-performing group of CIOs, which it calls “vanguards”.
The group comprises technology leaders who deliver better business outcomes and report high levels of AI value as perceived by their chief executives.
Companies with rules for dropping underperforming AI projects already in production were 2.6 times more likely to belong to this group.
Yet only 31% of CIOs said their organisations had criteria for stopping AI pilots, while 17% had guidelines for ending projects that failed to deliver value after deployment.
The gap comes as businesses move beyond initial trials of generative AI and traditional AI to deploy AI agents across their operations.
At the same time, CIOs face pressure to justify the cost of these investments.
Some 58% of respondents said they were under greater pressure to deliver cost savings through AI, while 51% expected AI to increase the total cost of owning and operating technology solutions.
Mullery said companies could no longer rely solely on funding and governance models designed for conventional technology investments, such as cloud computing and software-as-a-service products.
AI costs can fluctuate depending on usage and the resources required to run applications, making it harder for companies to predict expenses, Gartner said.
CIOs will need to track costs throughout a project’s life rather than focus only on the initial purchase.
Workforce planning is another area where companies are lagging.
Organisations with explicit plans to redesign jobs around teams of human workers and AI agents were 2.4 times more likely to be classified as vanguards.
Only 4% of CIOs, however, said their organisations had such plans.
AI adoption is not necessarily translating into fewer technology jobs. About 40% of respondents expected their IT headcount to increase in 2027, with cybersecurity and data and analytics among the areas expected to see growth.
As AI takes over specific tasks, companies may need fewer specialised job categories while expecting employees to handle a wider range of responsibilities.
Designers, for example, could take on testing work, while testers become more involved in design.
The findings suggest that companies seeking greater returns from AI will need to make changes beyond technology spending, including setting clearer investment priorities, controlling costs and rethinking how their teams work.
