Enterprise AI Faces a Harder ROI Test as Budgets Miss Forecasts

Enterprise AI discussions have shifted over the past 12 months as boards and CFOs press for clearer answers on implementation costs and bottom-line return on investment. [1] Gartner projects global AI spending by enterprises and vendors will reach $2.59 trillion, increasing the pressure on executives to scrutinize how initiatives are scoped, budgeted, and governed. [1] A survey of 500 senior finance leaders in the United States and United Kingdom found that 79% of large enterprises missed their AI budgets during the past 12 months. [1]

Another survey found that 85% of enterprises missed AI forecasts by more than 10%, while one quarter missed them by 50% or more. [1] At the same time, 6% of large enterprises reported EBIT impact above 5% from enterprise-wide AI implementation. [1] The article argues that ROI improves by defining and amplifying business value while defining and optimizing implementation cost. [1]

The contrast between broad forecasting failures and the 6% reporting EBIT gains above 5% suggests that AI ROI is not simply a deployment question; it is a selection, cost, and governance question. The article's numerator-and-denominator framing gives executives a practical test: reject initiatives whose business value remains vague, and scrutinize implementation costs before scaling. That approach also makes budget accuracy part of AI performance rather than a separate finance exercise. What the sources don't address: which specific governance controls, use cases, or cost categories distinguish the enterprises reporting stronger EBIT impact.