Uber has begun handing ChatGPT Work to more than 30,000 employees, OpenAI’s agent for office tasks, after testing the product with 1,000 staff. The company-wide deployment, announced Thursday, is Uber’s first agentic platform rolled out across its global workforce for knowledge work.
The move puts a general-purpose AI agent in the hands of operations, finance, marketing and product teams at once. Software engineers will keep using OpenAI’s Codex for coding, a sign that Uber is separating coding tools from the broader office suite.
What ChatGPT Work does
ChatGPT Work runs inside ChatGPT Enterprise and connects to Gmail, Google Drive, Google Calendar and Slack. It can research a topic, prepare a meeting brief, build a presentation, analyze data and carry a multi-step workflow through to a finished document. OpenAI launched the product in July.
An agent differs from a standard chatbot because it can plan several steps, use connected apps and return a finished artifact. That makes it more useful for recurring office work, but also harder to evaluate: a polished slide deck can hide weak analysis.
Uber will use GPT-6.1 Sol as the default model. OpenAI released that model in September and positioned it as a lower-cost option; TNW reported that it costs about one-fifth as much as the company’s top model. The choice matters because an agent can burn through far more tokens than a chatbot when it reads files, calls other tools and checks its own work.
The pilot was not limited to engineers. Sumanth Sukumar, Uber’s vice president of engineering, told TechCrunch that roughly 80% of the first 1,000 users came from outside engineering. They examined operational data for trends, adapted training material for local markets and assembled dashboards for daily tasks.
OpenAI says the rollout includes an internal Uber library of AI skills, approved integrations and training for employees. That support layer matters because access alone rarely changes how people work; staff need examples that fit their own tools and approval rules.
Why Uber is widening access
Uber is leaving room for teams to find their own uses. Sukumar said the company wants to put capable AI tools in place and let groups experiment, while it watches which workflows people build and whether the work improves. Uber has not disclosed what the rollout will cost.
The company has already tried a more hands-on version of that idea. Earlier this year it placed 30 AI engineers alongside teams for two weeks, asking them to automate parts of those teams’ jobs. Sukumar said the exercise showed the value of pairing technical skill with deep knowledge of a particular role.
Cost control is part of the backdrop. Uber’s chief technology officer, Praveen Neppalli Naga, said in August that the period of “tokenmaxxing,” or spending without limits on AI, was coming to an end, according to TNW. The publication also reported that Uber had exhausted its Claude Code allocation for 2026 by April.
Uber has also changed its workforce while expanding AI use. In July it cut 10% of its customer-service staff and, for the first time, tied the reduction to AI. The company has not said the ChatGPT Work rollout will eliminate roles, and its public framing stresses giving existing employees better tools.
OpenAI is using the deal to show that its enterprise business can scale beyond small teams. The company has pushed ChatGPT Enterprise and agent products as it competes with Anthropic and Google for corporate customers. A customer as visible as Uber gives its sales teams a reference that is easier to explain than a benchmark score.
What to watch
OpenAI and Uber say they will publish what they learn about adoption, common use cases and results as the deployment expands. Those details will be a better test than the size of the rollout. A rollout of this size gives OpenAI a prominent reference customer, but it does not by itself show that agents save time or raise the quality of work.
For Uber, the question is whether ChatGPT Work becomes part of daily routines or remains a tool people open occasionally. The company’s AI pods suggest that the largest gains may come from redesigning how a team works, not simply handing out more licenses. If Uber can show that difference across 30,000 people, other large employers will have a detailed case to examine.
Brian Shannon is just getting his start a reporter. He attended a technical school while still in high school where he learned a variety of skills, from photography to car mechanics. Brian also helps keep Techno Secrets social media feeds up-to-date.
