Gartner Finds CIOs Need AI Stop Rules

At Gartner IT Symposium/Xpo 2026 in Orlando on Tuesday, vice president analyst Alicia Mullery said CIOs achieving better business outcomes from AI deployments are more likely to set stopping points at both the pilot and production stages. Gartner found that just 11% of 2,002 CIOs and IT executives surveyed are currently delivering business outcomes while demonstrating AI value. Mullery said they are shifting from scaling everything to scaling value by quitting lower-value efforts and redirecting resources.

Gartner found that CIOs with guidelines for stopping an AI pilot before production are 2.3 times more likely to belong to that 11%. Those with policies for abandoning low-value AI initiatives already in production are 2.6 times more likely to be in the group. Only 31% of respondents had pilot-stage guidelines, and 17% had production-stage policies. Organizations averaged 10 AI use cases in 2024; Mullery said the average rose to 42 a year later, more than quadrupling.

Almost 60% of survey respondents reported increased pressure to deliver cost improvements from AI, while more than half expect AI to increase technology's total cost of ownership. Mullery said organizations should budget for AI across its full technology lifecycle, rather than treating it like traditional cloud and SaaS spending focused on the initial purchase. North American CIOs expect IT budgets to rise 2.9% in 2027, but Mullery said inflation may make the increase feel like a decline.

Mullery urged CIOs to work with CFOs in the coming year to create rules for AI economics, including assigning teams responsibility for token costs and seeking vendor transparency about full lifecycle costs. She cited Citigroup and Bank of New York Mellon as organizations using precise measures to manage token expenses; in some cases, they hold weekly executive governance meetings about those investments.

At Gartner IT Symposium/Xpo 2026 in Orlando on Tuesday, vice president analyst Alicia Mullery said CIOs achieving better business outcomes from AI deployments are more likely to set stopping points at both the pilot and production stages. Gartner found that just 11% of 2,002 CIOs and IT executives surveyed are currently delivering business outcomes while demonstrating AI value. Mullery said they are shifting from scaling everything to scaling value by quitting lower-value efforts and redirecting resources.

Gartner found that CIOs with guidelines for stopping an AI pilot before production are 2.3 times more likely to belong to that 11%. Those with policies for abandoning low-value AI initiatives already in production are 2.6 times more likely to be in the group. Only 31% of respondents had pilot-stage guidelines, and 17% had production-stage policies. Organizations averaged 10 AI use cases in 2024; Mullery said the average rose to 42 a year later, more than quadrupling.

Almost 60% of survey respondents reported increased pressure to deliver cost improvements from AI, while more than half expect AI to increase technology's total cost of ownership. Mullery said organizations should budget for AI across its full technology lifecycle, rather than treating it like traditional cloud and SaaS spending focused on the initial purchase. North American CIOs expect IT budgets to rise 2.9% in 2027, but Mullery said inflation may make the increase feel like a decline.

Mullery urged CIOs to work with CFOs in the coming year to create rules for AI economics, including assigning teams responsibility for token costs and seeking vendor transparency about full lifecycle costs. She cited Citigroup and Bank of New York Mellon as organizations using precise measures to manage token expenses; in some cases, they hold weekly executive governance meetings about those investments.