While many companies preach AI efficiency, Uber has turned its focus inward, automating core business workflows that were previously bottlenecks.
On The AI Daily Brief, Nathaniel Whittemore detailed Uber’s program of deploying agentic pods - teams that pair AI-proficient engineers with domain experts in finance, legal, and HR. Instead of analyzing process diagrams, these pods shadow daily work for two weeks. Within ten days, they ship a working agent for a specific, complex workflow.
"The real value isn't just the 20% efficiency gain on a single task. It's the total redesign of the workflow that eliminates handoffs and legacy tooling."
- Nathaniel Whittemore, The AI Daily Brief
The results are drastic. Uber CTO Praveen Napali reports that financial pacing reports, which previously took two days, now take ten minutes. Capital allocation decisions across 150 cities dropped from 15 hours to 30 minutes.
This marks a departure from the standard “AI champion” model, which Whittemore critiques as internal PR. The pods are a practical implementation focused on what AI can enable, not just preach. With 99% of Uber engineers using AI tools and over 70% of pull requests attributed to agents, the company is applying automation to the tasks that govern its business, not just peripheral ones.
The goal is to reinvest the massive time gains into work that was previously impossible to attempt. This suggests a model for enterprise AI adoption where the reward isn't just saved minutes, but unlocked ambition.
