Uber Uses OpenAI to Help Drivers Earn Smarter and Riders Book Faster
Explore integrating OpenAI Realtime API for voice features in your app to reduce friction and improve accessibility.
Review Uber's implementation of OpenAI Realtime API for voice and consider adopting a multi‑agent architecture for your own AI assistants.
Summary
Uber has partnered with OpenAI to launch AI‑powered assistants and voice features that help drivers earn smarter and riders book faster across its global marketplace. The company operates 40 million trips per day with 10 million drivers and couriers in 15 000 cities across more than 70 countries, requiring real‑time decision support. Uber Assistant, built on a multi‑agent architecture, routes each driver query to the most suitable model—using lightweight nano/mini models for quick classification and larger reasoning models for complex tasks. An internal AI Guard layer screens prompts and responses to enforce safety, privacy, and policy compliance, reducing hallucinations and maintaining consistency. Voice interactions are powered by OpenAI’s Realtime API, allowing users to request rides with natural speech while the system synchronizes spoken and visual responses. The beta rollout has already reached hundreds of thousands of U.S. drivers, who report faster ramp‑up and higher engagement. Uber’s engineering teams now embed AI across product, legal, operations, and design, accelerating experimentation and iteration cycles. This collaboration demonstrates how frontier LLMs can be integrated at scale to improve accessibility and productivity in a high‑volume marketplace.
Key changes
- Uber Assistant uses a multi‑agent architecture routing queries to specialized nano/mini or larger reasoning models
- AI Guard screens prompts and responses for safety, privacy, and policy compliance, reducing hallucinations
- Voice requests use OpenAI Realtime API to interpret intent and synchronize spoken and visual responses
- Uber Assistant beta has been rolled out to hundreds of thousands of U.S. drivers
- Uber’s engineering teams embed AI across product, legal, operations, and design, accelerating experimentation
- The partnership enables faster product iteration cycles through model specialization and continuous evaluation