McKinsey: agentic AI scales across companies, but the impact on profits fails to take off

By McKinsey & Company · September 2026.

In this mailing, Dan Tinkoff and Lieven Van der Veken, global leaders of QuantumBlack, AI by McKinsey, present the latest findings from the firm's State of AI research on enterprise adoption of artificial intelligence.

According to the study, 40% of respondents from large organizations say they are already scaling AI agents, up from 27% the previous year. In addition, about a third of companies are choosing to use agentic coding tools to build software products or features in-house.

Yet despite this progress in adoption, economic impact has not kept pace: the share of respondents reporting enterprise-wide EBIT impact from AI has held steady at 37%. In other words, employees who use AI see individual productivity gains, but for most companies those gains have yet to translate into profit growth.

The authors of the analysis (Dan Tinkoff, Lieven Van der Veken, Michael Chui and Tara Balakrishnan) break down in the linked content what sets apart the organizations that are capturing real value from AI from those that are not.

The email also promotes an event, "McKinsey Live: Improving the economics of agentic AI: A guide for business leaders", scheduled for September 22 and hosted by senior partners Tanguy Catlin and Lari Hämäläinen. The event description offers another notable figure: only 6% of companies see significant EBIT impact from AI, which makes closing the gap between spending and impact a critical governance challenge. The session promises to help business leaders weigh the real cost of the work agentic AI replaces against its value, and decide where to scale up and where to pull back.