Uber Expands OpenAI Partnership to Give 30,000 Employees Access to ChatGPT Work

Uber is expanding its partnership with OpenAI to provide ChatGPT Work to its entire global workforce of approximately 30,000 employees, extending an initial pilot as the ride-hailing company explores how generative artificial intelligence can improve productivity across its operations.

The expansion marks a significant increase from the 1,000 employees who participated in Uber's initial trial earlier this year. The pilot covered multiple departments, including finance, operations and product development, suggesting that the company sees potential applications for AI beyond software engineering and other technical functions.

Uber told TechCrunch that OpenAI worked closely with the company to integrate the chatbot into its operations. However, the ride-hailing giant did not disclose the cost of the expanded deployment or provide a specific estimate of the productivity gains it expects.

The decision to make ChatGPT Work available across the organization signals a shift from testing AI in selected teams to allowing employees throughout the business to identify and develop their own applications.

ChatGPT Work, introduced in July as an enterprise-focused AI agent, is designed to help corporate teams handle a range of workplace tasks. For Uber, the initial pilot provided evidence that the technology could be useful to employees whose responsibilities do not involve writing code.

Sumanth Sukumar, Uber's vice president of engineering, told TechCrunch that approximately 80% of the early adopters were non-engineers.

"They used it for a wide range of tasks: analyzing complex data from our operations to identify trends and develop recommendations; localizing training materials for different markets and languages; and building personalized interactive dashboards to manage their daily work," Sukumar said.

These applications point to several potential sources of value for Uber. Analyzing operational data could help teams identify patterns and develop recommendations, while translating and adapting training materials could reduce the time required to prepare resources for different markets. Interactive dashboards could also help employees organize information and monitor routine activities.

The examples illustrate how enterprise AI can be deployed across administrative, analytical, and operational functions rather than being confined to specialist technical teams. However, the company has not disclosed how much time these applications have saved or whether they have generated measurable cost reductions.

Despite expanding access to ChatGPT Work, Uber has not prescribed a single set of tasks that all employees must automate. Instead, the company intends to give teams flexibility to experiment with the technology and identify workflows suited to their specific responsibilities.

"Our approach is to make powerful AI tools available across Uber while giving teams room to experiment," Sukumar said.

The company has already tested a similar approach through targeted collaboration between technical specialists and other departments. Earlier this year, Uber paired 30 AI engineers with teams across the business and gave them two weeks to automate key parts of their jobs.

Those initiatives, which Uber called "AI pods", were designed to combine technical expertise with employees' understanding of the tasks they perform. The approach could help identify practical uses for AI more quickly than relying exclusively on centralized technology teams to design solutions for the entire organization.

"Those 'AI pods' showed us how much opportunity there is when you combine technical expertise with a deep understanding of a particular job," Sukumar said.

The model emerges from a broader challenge facing companies adopting generative AI: access to a capable model does not automatically produce business value. Employees must identify suitable tasks, incorporate the technology into existing processes, and determine when its output requires human review.

Giving workers room to experiment could reveal applications that senior management might otherwise overlook. At the same time, decentralized experimentation can make it harder to measure results consistently or ensure that different teams follow the same standards for accuracy, data protection and oversight.

Uber has not detailed how it will address those issues as adoption expands. Its stated plan is to monitor how employees use the technology, which workflows they develop, and whether those applications help them work more effectively.

"We'll look at how employees are using the technology, the kinds of workflows they develop and how it helps them work more effectively," Sukumar said. "We'll evaluate those outcomes and use what we learn to guide how we deploy AI across the business."

That evaluation will be important in determining whether the rollout becomes a source of lasting operational improvement rather than simply another enterprise software expense. Uber has not disclosed the financial terms of the agreement, making it difficult to assess the investment against any expected savings or productivity gains.

The expanded agreement comes as OpenAI accelerates its efforts to build its enterprise business and compete with Anthropic for corporate customers. Although ChatGPT initially gained widespread recognition as a consumer product, businesses have increasingly become an important market for AI providers seeking recurring revenue from workplace deployments.

Enterprise adoption offers a different commercial opportunity from individual subscriptions. A company-wide agreement can place an AI tool within the daily workflows of thousands of employees, potentially encouraging sustained usage across departments. But large deployments also require providers to demonstrate reliability, security, and practical value, particularly when customers are deciding whether to renew or expand their contracts.

The competitive stakes are increasing as businesses assess products from OpenAI, Anthropic and other AI developers. The ability to win enterprise customers is not determined solely by model performance; integration, employee adoption, governance and demonstrable returns can also influence purchasing decisions.

Therefore, Uber's rollout marks a large-scale test of whether general-purpose AI can improve productivity across a complex business with varied operational needs. The number of employees given access establishes the potential reach of the deployment, but it does not by itself demonstrate its success.

The more consequential measure will be what Uber learns from the rollout: which workflows employees adopt, how reliably those workflows perform, and whether the resulting improvements justify the cost. Those outcomes will help determine whether broader AI access becomes a meaningful source of efficiency for Uber and a stronger enterprise growth opportunity for OpenAI.