OpenAI published two studies on August 12, 2026 that track how companies are using artificial intelligence inside their operations, reporting that enterprise usage is moving from answering questions toward completing work.
The company measured output tokens per active user across its enterprise customer base and found a widening gap between heavy adopters and everyone else. Firms in the top 10% of monthly usage, which OpenAI calls frontier firms, generated 8.3 times as many output tokens per active user as firms sitting between the 45th and 55th percentiles. That gap stood at 2.6 times in January.
Agent tools account for a growing share of that volume. As of June, Codex produced 64% of the combined Codex and ChatGPT output tokens among enterprise customers. OpenAI attributes the figure to agentic workflows running longer multi-step tasks that generate more output than a single chat exchange.
Advanced features separate the two groups. Among weekly active users, 21% at frontier firms use Plugins and 19% use skills, compared with 9% and 3% at typical firms.
Growth in agent use has spread well past software teams. Since February, weekly active enterprise Codex users rose 108 times in legal, 41 times in sales, 41 times in recruiting, and 26 times in marketing, against 5 times in engineering.
The companion working paper examined US public companies and found that enterprise adopters held more assets, employed more workers, and reported higher research and development spending than non-adopters. Usage also skewed young. Six months after adoption, early-career employees sent 13 more messages per week than executives.
Source: OpenAI - https://openai.com/index/how-enterprises-put-ai-to-work/