Uber and OpenAI said Thursday that they are expanding their partnership, with the ride-hailing company putting its entire global workforce of roughly 30,000 people on ChatGPT Work. The move turns a limited experiment into one of the larger known enterprise deployments of OpenAI's business-focused AI agent, and gives both companies a high-profile reference case as corporate adoption of generative AI shifts from pilots to company-wide rollouts.

From a 1,000-person pilot to the whole company

ChatGPT Work launched in July as an enterprise-oriented agent that corporate teams can deploy across workflows. Uber began testing it earlier this year with about 1,000 employees drawn from departments including finance, operations and product. That pilot produced a notable signal: roughly 80% of the early adopters were non-engineers, according to Sumanth Sukumar, Uber's VP of engineering.

The usage patterns in the pilot were broad rather than narrowly technical. Sukumar said participants analyzed complex operational data to spot trends and build recommendations, localized training materials for different markets and languages, and built personalized interactive dashboards to manage their day-to-day work. That profile matters for enterprise architects: it suggests the value of the tool is landing in knowledge work adjacent to core engineering, where data wrangling and content adaptation are routine but time-consuming.

Experimentation, not a fixed playbook

Uber is going all-in on ChatGPT Work without having settled exactly how employees will use it. "Our approach is to make powerful AI tools available across Uber while giving teams room to experiment," Sukumar said. He pointed to an earlier internal effort in which 30 AI engineers were paired with teams across the company and given two weeks to automate key parts of their jobs. Those "AI pods," as Uber calls them, demonstrated the opportunity that emerges when technical expertise is combined with deep knowledge of a specific job.

The company did not disclose what the expanded rollout will cost. Instead, Sukumar said Uber will track how employees use the tool, the kinds of workflows they develop and how it helps them work more effectively, then use those outcomes to guide how AI is deployed across the business. OpenAI worked closely with Uber on integrating the chatbot, according to the company.

An enterprise land grab

For OpenAI, the Uber deal is part of a rapid push to scale its enterprise business as it competes with Anthropic. Once known largely as a consumer brand, the ChatGPT creator has moved quickly to court business customers, and large deployments like this one serve as both revenue and proof that its enterprise offerings can operate at workforce scale. The competitive pressure is mutual: Anthropic has made enterprise and coding use cases central to its own positioning, and the two companies are now competing for the same corporate budgets.

The technical questions that follow a rollout of this size are familiar to IT leaders. Access control and data governance become more complex when tens of thousands of employees across markets and languages use an AI agent that touches operational data. Integration with internal systems, observability into how the tool is used, and cost management at scale all become operational concerns rather than pilot-stage curiosities.

What to watch

Uber's decision to measure outcomes rather than prescribe usage is a pragmatic approach for a company that spans mobility, delivery and freight operations across many markets. Whether that experimentation produces durable productivity gains — and whether the workflows employees build can be standardized and governed — will determine if the rollout becomes a template for other large enterprises. For now, Uber has moved generative AI from a departmental trial to a default tool for its global workforce, and the results of that bet will be watched closely across the industry.