Uber brings ChatGPT Work to all employees with OpenAI

Uber is rolling out ChatGPT Work, OpenAI's agent inside ChatGPT Enterprise, to more than 30,000 employees, expanding from a 1,000-person pilot. OpenAI describes it as Uber's first company-wide deployment of an agentic platform for knowledge work, built to take on multi-step tasks and deliver finished work.

The agent plugs into tools like Google Drive, Gmail, Google Calendar and Slack to research, prepare meeting briefs, build presentations, analyze data and run workflows across apps. GPT-6.1 Sol is the default model. This is a large employer moving agents from experiment to everyday infrastructure.

We think this is the kind of deployment the industry needs more of, because it will produce evidence. Pilots are easy to love. A rollout across operations, finance, marketing and product at a company of Uber's size is where agents either save real time or create a lot of confident busywork. Both companies say they'll share insights on adoption, use cases and measurable outcomes as it progresses. We'll hold them to that, and we'd like to see the misses alongside the wins.

There's a lot to like in how Uber is framing it. The rollout comes with an internal AI skills library, approved integrations and dedicated employee training. That's the unsexy scaffolding most enterprise AI rollouts skip, and it's usually why they stall. Engineers, meanwhile, keep using Codex for software work, so this is about extending agents to everyone else.

Now the harder questions. An agent with access to email, calendars, shared drives and Slack is an agent with access to a company's most sensitive material: personnel matters, finances, strategy, customer issues. Connecting all of that to an autonomous system that acts across applications raises the stakes on permissions, logging and mistakes. A meeting brief that misstates a number is annoying. An agent that sends, shares or edits the wrong thing at scale is a governance incident.

There's also the workforce question we won't skip. Uber frames this as giving everyone tools to innovate, and we take that at face value. But when an agent delivers "finished work" on presentations and analysis, that's the work junior staff used to do to learn the job. If this becomes the norm across large employers, the pipeline for building expertise thins, and companies will need to be deliberate about how people still learn.

Our read: a credible, well-supported deployment and a real test of agentic AI at work. The proof will be measured outcomes, clean security, and whether employees end up doing more meaningful work or just reviewing more machine output.

OpenAI announced the rollout October 22, 2026, with comments from Uber CTO Praveen Neppalli Naga and OpenAI president Greg Brockman. ChatGPT Work also gives Uber access to OpenAI's multimodal tools, including voice and image generation, alongside its frontier models.