Uber cut 3,300 jobs in September. Now it is giving 30,000 employees ChatGPT

Uber's workforce changes and ChatGPT deployment

On September 2, Uber told employees it was eliminating about 10% of its corporate workforce, roughly 3,300 jobs, to simplify the organization and strip out management layers. On October 22, OpenAI announced that ChatGPT Work is going to more than 30,000 Uber employees worldwide, the company's first company-wide deployment of an agentic platform for knowledge work, according to Tech Times UK. Seven weeks separate the two announcements. Neither company has said they are connected, and Tech Times UK reports there is no evidence the cuts were driven by the tool. The sequence still tells its own story about how a large company is choosing to run.

The rollout expands a pilot of about 1,000 employees. ChatGPT Work goes to operations, finance, marketing and product teams, while software engineers keep using Codex, OpenAI's coding tool, Tech Times reported. The platform connects to Google Drive, Gmail, Google Calendar and Slack, and Uber is backing it with an internal library of AI skills, a set of approved integrations and dedicated training. OpenAI described the intended uses as research, meeting briefs, presentations, data analysis and workflows that cross applications.

Lessons from the pilot and future plans

The pilot's lessons are the practical part. Uber vice president of engineering Sumanth Sukumar told TechCrunch that about 80% of early adopters were not engineers, and that the company paired 30 AI engineers with teams for two weeks in what it called AI pods to automate pieces of their work, Tech Times UK reported. Examples from the pilot included analyzing operational data, localizing training materials and building personalized dashboards. Uber's chief technology officer, Praveen Neppalli Naga, said in the announcement that he wants to see what teams build with it. OpenAI president Greg Brockman described the deployment as a step toward putting advanced tools in the hands of more employees for complex work.

Financial transparency and measuring outcomes

What is missing is the money and the measurement. Uber did not disclose what the deployment costs, per TechCrunch, and OpenAI's announcement included no productivity or time-saving figures. Both companies say they will share adoption and outcome data later, and Uber plans to track usage, workflows and effectiveness as the rollout proceeds.

Disconnect between layoffs and AI rollout

The September cuts were presented as structure, not technology. Chief executive Dara Khosrowshahi said the restructuring was meant to reduce management layers, including teams where managers had as few as one or two direct reports, and he did not cite AI in the memo announcing them, according to Business Insider's reporting on the layoffs. Uber described the changes as organizational and not a judgment on anyone's contribution.

The people who lost their jobs saw it differently. Business Insider interviewed laid-off Uber employees in September who described AI use growing rapidly in their roles before the cuts. One said a tool he had been using in the weeks before his layoff had reduced the time his job took to about a fifth of what it had been. Managers in the advertising unit had been encouraging staff to use Gemini or ChatGPT to prepare for customer meetings. Uber has also, in a separate action, reduced its customer service team by about 10% in a move Engadget reported was explicitly tied to AI, and the company had previously said it was slowing hiring because of AI.

Put together, the picture is a company that denies a causal link between the layoffs and the AI rollout while making decisions that only make sense if one informs the other. Flatten the org first; equip the people who remain second. American Express gave shareholders the same sequence this week, cutting 5% of its staff while describing the move as preparation for automation.

A template for software and technology buyers

For the software and technology buyers watching, Uber's deployment is a useful template because of what it includes beyond licenses: a pilot that measured who actually used the tool, engineers embedded with business teams to build the first workflows, a curated set of integrations rather than an open door, and a commitment to publish results. The license count makes the headline.

The open question is the one both companies have deferred. If ChatGPT Work makes a 30,000-person company materially more productive, the next restructuring memo will be harder to write without mentioning it.