Uber Rolls Out ChatGPT Work to All 30,000 Employees
Uber will deploy ChatGPT Work to its entire global workforce of about 30,000 people, according to an announcement from Uber and OpenAI on October 22, 2026. The companies described the move as an expansion of their existing partnership. TechCrunch reported the rollout the same day.
ChatGPT Work is an enterprise-focused AI agent built for corporate team workflows. An agent differs from a chat assistant in that it can take actions in connected systems, such as updating a spreadsheet or filing a request, rather than only producing text for a person to act on. Uber first piloted it with 1,000 employees earlier in 2026, including staff from finance, operations, and product departments.
About 80% of Uber's early ChatGPT Work adopters were non-engineers. That share matters because it places the tool with employees whose daily work runs through spreadsheets, internal tools, planning documents, and cross-team requests rather than code repositories.
To accelerate adoption, Uber paired 30 AI engineers with teams across the company for two weeks. The engineers worked with those teams to automate key parts of their jobs. OpenAI published a company news post on October 22, 2026, titled "Uber brings ChatGPT Work to all employees with OpenAI," which OpenAI's news page lists among its company announcements.
In my view, the staffing model is the more instructive detail. Handing out licenses, which are typically priced per user seat, is a distribution task. Assigning engineers to sit with business teams for two weeks treats adoption as a redesign of how the work gets done. Enterprises that have watched seat-based rollouts stall after the first wave of logins will recognize the difference.
The broader context here is the operational groundwork a deployment of this scale requires. An agent working across finance, operations, and product touches permissions, data handling, and audit trails, meaning the records of what a system did and why. Organizations considering a similar step will need to settle how agent actions are logged, which systems an agent can reach, and who reviews its output before it affects a decision. Those questions apply regardless of vendor.
The practical upside, as I read it, is that the people closest to recurring work are now shaping how it gets automated. If the pilot's gains carry across a workforce this size, the change will show up less in headline productivity claims than in hours returned to tasks non-engineering teams have long handled by hand. That outcome depends on execution, and the coming months of rollout will show how far the pilot results travel.
