Codex Economic Impact Study Shows Rapid Adoption

The Codex economic research paper quantifies the rapid adoption of Codex across individual, organizational, and internal OpenAI users, showing a dramatic shift from chatbots to agentic tools as the primary AI interface.

By May 2026, 80.6% of individual users made Codex requests >30 min, 70.2% >1 hour, and 25.6% >8 hours. Codex became the primary AI tool for all OpenAI departments, >85% output tokens. Non‑developer adoption grew 137× (individual), 189× (org), 12× (internal). Codex enables non‑technical teams to perform tasks requiring technical expertise. Users orchestrate multiple agent tasks per day, 99th percentile >60 hours of agent turns. Findings suggest shifting to agentic AI for long‑horizon, cross‑functional tasks. The study demonstrates that agentic tools can lower the cost of moving across task boundaries and expand the frontier of productive work for businesses. Codex’s economic impact is evident in the increased token usage and deepened engagement across departments. The findings suggest that organizations should consider shifting to agentic AI to handle long‑horizon, cross‑functional tasks. The research underscores the importance of integrating agentic tools into workforce strategies to maximize productivity gains.