ORLANDO, FLA. — CIOs who’re reaching the very best outcomes from their AI deployments are those who know when to press pause.
That may appear counterintuitive, given the tempo at which AI is evolving. However Gartner’s current CIO survey recommended that IT executives who’re delivering on higher enterprise outcomes whereas proving AI worth are extra doubtless than their friends to ascertain two stopping factors within the AI lifecycle — on the pilot and manufacturing levels.
“They’re shifting from simply scaling all the pieces to scaling worth, which suggests they’ve a really robust self-discipline of quitting and redirecting assets to what’s extra priceless,” stated Alicia Mullery, vp analyst on the consultancy, throughout a presentation Tuesday at Gartner IT Symposium/Xpo 2026.
There aren’t a lot of them. Gartner stated simply 11% of the two,002 CIOs and IT executives who participated in its annual survey are at the moment threading that needle.
But Gartner discovered that CIOs who’ve established pointers for when an AI pilot undertaking should not transfer into manufacturing are 2.3 instances extra prone to be a part of that 11%; likewise, CIOs who’ve established insurance policies on when to desert low-value AI initiatives already in manufacturing are 2.6 instances extra prone to be among the many elite group.
Nonetheless, solely 31% and 17% of CIOs, respectively, stated they’ve established some of these insurance policies and pointers, even because the stakes get larger.
In 2024, organizations juggled on common 10 AI use circumstances; a yr later, that quantity greater than quadrupled to 42 use circumstances , in line with Mullery.
“We’re getting superb at launching AI,” she stated. “We’re not good at stopping AI.”
Certainly, AI budgets are rising, however so is the stress for CIOs to display ROI. Nearly 60% of survey takers reported an uptick in stress to ship value enchancment from AI. But greater than half predict that AI will enhance the entire value of possession of know-how.
To raised monitor these will increase, CIOs want to regulate how they construct their AI budgets. Mullery defined that many organizations funds for AI the identical means they’d conventional know-how investments, together with cloud and SaaS, when they should take into account the price of AI over the know-how’s full lifecycle, not simply the preliminary know-how buy.
Understanding when to pause AI deployments will probably be crucial for CIOs as they plan for subsequent yr’s budgets. Whereas CIOs in North America count on a 2.9% IT funds enhance in 2027 , Mullery famous that this enhance will doubtless really feel like a funds decline attributable to inflation.
To raised handle AI budgets, Mullery inspired CIOs to work with CFOs to “co-create new guidelines of AI economics” within the coming yr. These new guidelines embrace figuring out which groups are liable for token prices and acquiring transparency concerning the full lifecycle value of AI from distributors, for instance.
“Organizations like Citigroup, like Financial institution of New York Mellon, particularly known as out how they’re placing in exact measures on easy methods to handle these token prices, and in some circumstances, they’re placing in weekly government governance for easy methods to handle these investments,” Mullery stated.
Are you an elite CIO? Tell us at [email protected].
