All posts
2 min readby Romiel Inolino

Gartner: only 13% of 2,002 CIOs and tech leaders report major returns from AI. The winners decide early what to shut down

AI ROIGartnerAutomation strategyCIOBusiness automation

Most companies still cannot show real returns from AI, and Gartner's survey points at project choice, not model choice.

What happened

According to The Decoder's report on a Gartner survey of 2,002 CIOs and other technology leaders, presented at the Gartner IT Symposium in Orlando, only 13% report major or transformative returns from AI.

The pressure runs both ways. 58% of CIOs are tasked with delivering cost savings through AI, yet 51% expect AI to raise total operating costs for their tech stack. 44% of CEOs struggle to show AI-driven financial returns to stakeholders, and 75% of CEOs still plan to increase technology spending.

Gartner analyst Alicia Mullery described a gap between the "innovation race" vendors are running and the "outcomes race" companies need to win. The companies that succeed decide early which projects are worth pursuing and which to shut down, even ones already in production. They also plan ahead for how AI agents will change job roles.

My take

This matches what I see with small and mid-sized clients. The failed projects rarely failed because the model was weak. They failed because nobody wrote down what the automation was supposed to save, so nobody could say whether it worked.

Before I build anything, I want three numbers: how many times the task happens per week, how long it takes a person, and what an error costs. That gives a baseline the automation has to beat. After launch, the same numbers decide whether it stays, gets fixed or gets switched off.

The "shut it down" part is the hard one. A workflow that runs every day but saves ten minutes a month still costs API calls, maintenance and attention. Review your automations every quarter like a portfolio and retire the ones that do not earn their keep. Also watch the cost line: if half of CIOs expect AI to raise operating costs, price every model call into the business case from day one, not after the first invoice.

More posts