The State of Enterprise AI 2025: From Experimentation to Core Infrastructure
OpenAI’s 2025 Enterprise AI report provides a data-driven view into how large organizations are embedding AI into core workflows, where productivity gains are materializing, and why a widening gap is emerging between AI leaders and laggards.
OpenAI
Office of the Chief Economist, OpenAI
Table of contents
Share
Based on aggregated usage data from more than one million business customers and surveys of 9,000 enterprise workers, this report shows that enterprise AI adoption is accelerating rapidly—but unevenly. Organizations are moving from isolated pilots to deeper workflow integration, with ChatGPT Enterprise message volume growing 8x year-over-year and API reasoning token usage increasing more than 300x. Workers using AI report saving 40–60 minutes per day, and frontier users—those who embed AI across a wider range of tasks—realize disproportionately higher productivity gains. At the same time, a widening divide is emerging: frontier firms generate roughly twice the AI usage per seat and seven times the use of advanced tools compared to median enterprises. Case studies across industries such as finance, healthcare, retail, and technology demonstrate measurable impacts on revenue growth, customer experience, and cycle-time reduction. The report concludes that the primary constraint to AI value creation is no longer model capability, but organizational readiness—specifically leadership alignment, workflow redesign, data integration, and change management.



