Uber Deploys ChatGPT Work to All 30,000 Employees in Largest Enterprise AI Agent Rollout
Uber announced this week a significant expansion of its partnership with OpenAI: the mobility and logistics company is now distributing ChatGPT Work to its approximately 30,000 employees worldwide. The announcement, made on October 22, 2026, marks what OpenAI describes as the first large-scale corporate deployment of an AI agent platform designed for knowledge work.
The decision did not come overnight. Earlier in the year, Uber conducted a pilot with 1,000 employees across different departments — finance, operations, and product — to test ChatGPT Work under real conditions. The results were sufficient to justify immediate expansion to the entire workforce. According to Sumanth Sukumar, Uber's vice president of engineering, approximately 80% of early adopters were non-engineers, demonstrating that ChatGPT Work's impact concentrates on knowledge work roles that traditionally lack access to sophisticated automation tools.
What ChatGPT Work Does in Practice
ChatGPT Work is not simply ChatGPT with a different logo. Introduced in July 2026 as an enterprise tool, it operates as an autonomous agent within the ecosystem of applications employees already use daily. Uber's integration connects ChatGPT Work to Gmail, Google Drive, Google Calendar, and Slack — the same tools that form the backbone of the company's internal productivity.
The use cases that emerged from the pilot are emblematic. Operations employees analyzed complex data to identify trends and develop recommendations; human resources teams localized training materials for different markets and languages; product professionals created personalized interactive dashboards to manage daily work. The common thread among all these scenarios is that none of them can be solved with a simple question and answer — each requires multiple steps, access to different systems, and the ability to produce a final ready-to-use artifact.
OpenAI's description of the product is straightforward: ChatGPT Work "can execute complex multi-step tasks and return finished work." In practical terms, this means an employee can ask the agent to prepare a complete presentation, compile data from multiple sources, formulate recommendations, and deliver the result in a ready-to-use format — all without intermediate manual intervention.
The GPT-6.1 Sol Choice: Cost Efficiency as Strategy
A detail that deserves attention is the technical decision behind the implementation. ChatGPT Work at Uber runs, by default, on the GPT-6.1 Sol model — OpenAI's most recent model, released in September 2026, positioned to balance frontier performance with cost efficiency. According to The Next Web, GPT-6.1 Sol costs approximately one-fifth of OpenAI's most expensive model.
This choice is not merely technical; it's financial. Uber went through a process of reflection about its AI spending. In August 2026, CTO Praveen Neppalli Naga publicly declared that the era of "tokenmaxxing" — the period of unrestrained AI investment without rigorous ROI measurement — was drawing to a close. According to Reuters and specialized outlets, the company had already used its entire 2026 budget for Anthropic's Claude Code before even April, leading to a strategic reevaluation of which AI tools would be worth scaling.
The decision for OpenAI, therefore, is not just about model quality, but about cost-benefit. GPT-6.1 Sol offers sufficient capability for corporate tasks without the premium price of more advanced reasoning models, and OpenAI's multi-model architecture allows teams to choose the most appropriate model for each need.
Engineers Apart: Codex Remains the Core
Uber did not replace all AI tools with ChatGPT Work. Software engineering teams continue using Codex, OpenAI's coding tool, specifically designed for development. This separation reflects an industry truth: programming tasks require different contexts and integrations than knowledge work tasks. Codex operates in the development environment, while ChatGPT Work operates in everyday productivity tools.
The "AI Pods" Experiment
Before the full rollout, Uber conducted an internal experiment that illustrates how the company is thinking about AI integration into its culture. In the first months of 2026, the company paired 30 AI engineers with teams from different departments and gave them two weeks to automate the most critical parts of each team's work. The result — described by Sukumar as a demonstration of "how much potential exists when technical expertise combines with deep understanding of a specific job" — was decisive for justifying the expansion.
The concept of "AI pods" suggests that Uber is not merely delivering a tool and hoping employees will use it. There is a support infrastructure: an internal AI skills library, approved integrations, dedicated training, and a continuous evaluation process. Sukumar told TechCrunch that the company "will look at how employees use the technology, the types of workflows they develop, and how it helps them work more effectively."
The Broader Context: Uber, OpenAI, and the Enterprise AI War
Uber's expansion does not occur in a vacuum. OpenAI has been working to position itself as an enterprise company, competing directly with Anthropic and Microsoft Copilot. OpenAI CEO Sam Altman has repeatedly declared that the corporate market is the company's growth engine, and each large rollout like Uber's serves as proof-of-concept for potential customers.
The relationship between Uber and OpenAI was already documented before this announcement. In May 2026, OpenAI published a detailed case study detailing the use of its models for Uber's driver- and rider-facing products — including Uber Assistant, an AI assistant that transforms marketplace data like earning trends and heatmaps into positioning insights, and that now operates for hundreds of thousands of US drivers. Uber also developed AI Guard, an internal governance layer that screens prompts and responses to promote safety, privacy, and security, enforce policies, and reduce hallucinations.
Now, with ChatGPT Work, Uber extends the relationship beyond marketplace products and into its own workforce — a step that signals maturity in AI adoption.
What to Watch
Two aspects deserve tracking over the coming months. The first is the success metric: Uber chose not to disclose the financial value of the contract nor publish productivity measurements. Instead, the company will track adoption, developed workflows, and perceived effectiveness — an approach that can generate valuable insights, but also leaves the ROI question open.
The second aspect is the context of AI-linked layoffs. In July 2026, Uber laid off 10% of its customer service team, the first time the company directly linked staff cuts to AI implementation. The expansion of ChatGPT Work to 30,000 employees raises questions about how agent automation will transform — and perhaps reduce — the need for certain roles in the future.
The question is whether the "give everyone the tool and observe" model can truly capture the value of corporate AI, or whether success will require more rigorous metrics and a willingness to make deeper organizational changes than most companies are prepared to make.
[This article was published by MPDev. Sources: TechCrunch, The Next Web, Unite.AI
Sources: TechCrunch, The Next Web, Unite.AI
✓ Independent sources cross-checked and verified before publishing
