Uber and OpenAI expanded their partnership on October 22, 2026: nearly 30,000 employees of the company worldwide will get access to ChatGPT Work starting today. Uber disclosed this to TechCrunch on Thursday, and OpenAI issued an official announcement. Both companies frame the move as an expansion of their mutual partnership. The rollout begins on the announcement day: nearly 30,000 employees will get platform access from October 22. The expansion follows staged testing: the platform was first piloted with 1,000 employees, and now it opens to teams across all company functions.
What the Announcement Means
Nearly 30,000 Uber employees — the company's entire global workforce — will get access to ChatGPT Work. In OpenAI's description, this is the first full-scale deployment of an agentic knowledge-work platform at Uber's scale. This is the next step after the pilot: the platform is leaving its limited test scope and opening to all company employees.
In OpenAI's official announcement, the expansion is called "the first full-scale deployment of an agentic knowledge-work platform at Uber's scale"; the platform's job is to take on complex, multi-step assignments and deliver a finished work result.
An agentic platform is something different from a simple chat window: the system breaks a user's task into several steps, interacts with the necessary apps, and ultimately delivers finished work. That approach is at the center of Uber's expansion decision. The rollout is so large that after a 1,000-person pilot, the platform opens at once to the entire 30,000-person workforce — a single full transition was chosen instead of gradual expansion.
What the Platform Does
ChatGPT Work connects to four core tools Uber employees use every day: Google Drive, Gmail, Google Calendar, and Slack. Through these integrations, employees can search for information, prepare briefings for meetings, create presentations, analyze data, and complete workflows across multiple apps.
In practice, this means, for example, that meeting preparation no longer requires manually gathering separate documents: the agent sees the meeting in the calendar, finds the relevant files in Drive, accounts for correspondence in Gmail, and prepares a single briefing document. When preparing a presentation, it collects data from several sources and delivers it as slides — work is not copied manually from one app to another, but completed within a single workflow. In analytical tasks, it processes large volumes of operational data and presents conclusions as a presentation.
Inside the platform, GPT-6.1 Sol runs as the primary model. In addition, OpenAI's multimodal capabilities such as voice and image generation are unlocked — teams can choose the right model and tool for each need. Uber's developer and engineering teams will continue using Codex to speed up software development and simplify coding workflows.
What the Pilot Revealed
The expansion decision rests on the results of a pilot with 1,000 employees. Moving from 1,000 pilot participants to the entire 30,000-person workforce means roughly a thirtyfold expansion of coverage. Notably, about 80% of the early users were not engineers — they were from finance, operations, and product teams, Uber's vice president of engineering Sumanth Sukumar told TechCrunch.
According to them, employees used the platform in three main ways: analyzing operational data to identify trends and develop recommendations, adapting training materials to different markets and languages, and creating personal interactive dashboards to manage daily work. In other words, the tool was mainly tested on easing the daily workload of non-technical employees.
Before the expansion, Uber ran one more experiment: the company attached 30 AI engineers to various teams and organized two-week "AI pods" sprints. In these sprints, teams tried to automate core parts of their work — unlike the individual usage experience in the pilot, technical specialists here worked directly inside the teams.
Open Questions
Uber did not disclose how much it is spending on the new rollout — the project budget remains unknown. Instead, the company chose another path: it will track how employees use the technology, what workflows they create, and how this affects their productivity.
The results will be evaluated, and the findings will guide plans for further AI deployment across the business — which workflows employees create and how this affects their productivity will be studied precisely through these observations.
