OpenAI study finds top enterprise users generate 8.3 times more tokens as agentic AI spreads

What happened
An OpenAI research study on enterprise usage shows that global companies are transitioning from basic AI assistance to execution. According to the published research details, top-tier frontier firms generate 8.3 times more output tokens per user as automated agentic workflows spread across their corporate operations.
Why it matters
The significant increase in token output indicates that businesses adopting agentic workflows are moving beyond conversational drafting toward automated task execution. This shift implies that enterprise AI utilization expands as systems perform tasks independently, changing computational demands and operational processes within businesses.
Bigger picture
This trend highlights a broader technological shift from advisory software tools to enterprise-wide execution systems. As frontier firms lead the adoption of agentic AI, corporate competitiveness may increasingly depend on integrating automated workflows rather than relying solely on manual user prompts.
Watch next
Watch for further research findings from OpenAI, published metrics regarding enterprise return on investment, and corporate governance frameworks established for automated agentic execution.
