AI's promised savings aren't materializing: Just 3% of companies beat cost-cutting targets, but 85% plan to spend more
The figure comes from Gartner's HR Symposium in Orlando. In a finding that perfectly sums up the blind faith so many executives have in AI, 77% of board members expect the technology will deliver value in the form of productivity and cost savings. This is despite 97% of organizations being... Read Entire Article
Recent data from Gartner reveals that only a minuscule 3% of companies have successfully exceeded their cost-cutting targets using AI, despite 97% of organizations failing to achieve savings through this technology alone. Senior Director Analyst CV Viverito warned that organizations are unlikely to generate cost savings from AI at present. Instead, the greater opportunity lies in using AI to introduce new capabilities, outcomes, and growth sources, transforming employee time rather than merely saving it.
However, the promise of cost reduction has not deterred businesses from increasing their AI spending. In fact, 85% of companies indicate they plan to invest even more in AI next year. This trend echoes similar findings from PwC's Global CEO survey, which found that 56% of CEOs saw no cost or revenue benefits since implementing AI. Additionally, a 2025 MIT report revealed that 95% of AI implementations fail to increase profits, while a 2024 survey indicated that generative AI might even be hurting workplace productivity.
To address these challenges, Gartner suggests that companies first agree on what they aim to achieve with AI. Their survey of 297 chief HR officers found that 78% reported no consensus among leaders on the value AI creates. Gartner recommends focusing investments where AI can make a difference, helping managers redesign work, and utilizing existing AI expertise. Finally, the organization advises using the time saved by AI to promote innovative work, of which only 21% was reported.
Written by urgent.news from TechSpot's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.