Tech Executives Push AI ROI Metrics as Only 37% See Profit Impact
Updated
Updated · Computerworld · Oct 5
Tech Executives Push AI ROI Metrics as Only 37% See Profit Impact
1 articles · Updated · Computerworld · Oct 5
Summary
McKinsey’s latest AI survey shows the measurement gap driving executive scrutiny: about 80% of respondents reported productivity gains, but only 37% said AI improved profits and just 6% tied it to at least 5% of operating profit.
Analysts say better ROI is emerging from workflow redesign rather than simply adding agents to existing tasks, with CFOs pressing CIOs and investors pressing CEOs to show where AI value is actually landing.
Gartner warned costs are a major obstacle, with three in five IT leaders worried about AI agents generating unexpected bills; token spending for coding assistants can already exceed human developer costs.
That is pushing companies to track unit economics in pilots—such as cost per completed task—and to choose smaller or open-weight models when they deliver acceptable results at lower expense.
Broader ROI still depends on foundations beyond finance, analysts said, including data quality, governance and AI control layers; KPMG found 55% of organizations have a formal harness layer, rising to 86% among those reporting established ROI.