OpenAI’s push into agentic AI, the kind that doesn’t just answer questions but actually does things across your apps, has crossed the 10 million user mark. The milestone comes on the heels of ChatGPT Work’s launch on July 9, 2026, which positioned OpenAI squarely in the territory of enterprise workflow automation.

ChatGPT Work functions as an agent that pulls context from multiple applications to handle multi-step tasks.

From Codex to ChatGPT Work: the adoption ramp

At launch, ChatGPT Work inherited more than 5 million weekly users from Codex, OpenAI’s coding-focused tool, plus over 1 million non-development users who were already leaning on the platform for tasks beyond writing code.

By November 2025, ChatGPT for Work had exceeded 7 million seats. That represented a 40% increase in just two months.

Enterprise seats for ChatGPT Work grew 9x year-over-year as of November 2025. For context, most enterprise SaaS companies celebrate 30-40% annual growth.

Approximately 10 million ChatGPT Plus subscribers contributed to the overall ecosystem.

Workspace agents and the collaboration play

On April 22, 2026, OpenAI introduced workspace agents, a feature designed for teams rather than solo users. These agents operate within permission-controlled environments and handle shared automation of complex tasks.

This collaborative layer represents a natural evolution from the general-purpose ChatGPT Agent launched in July 2025, which was available to Pro, Plus, and Team subscribers.