Uber Says AI Agents Booked Just Four Rides in Its Busiest Week

AI agents that book rides and order food for you are one of the most hyped ideas in tech right now, but Uber’s own numbers suggest almost nobody is using them yet. On its third-quarter 2026 earnings call on November 4, Uber said only four rides were booked through AI agents during its busiest week of October.

What Uber reported

Uber’s chief financial officer, Balaji Krishnamurthy, shared the figures while discussing how customers reach the platform. During the week of Halloween, the company’s busiest week of the month, Uber handled roughly 320 million trips and orders across its ride-hailing and Uber Eats businesses.

Of those, just four ride bookings came from an AI agent, according to TechCrunch’s report on the call. Food delivery did slightly better: about 500 Uber Eats orders were placed by agents in the same week. Both numbers are tiny next to the company’s total volume.

The figures landed on the same day Uber reported strong overall results. The company said gross bookings rose 20% year over year to $59.8 billion in the quarter, and that trips grew to 4.1 billion. AI agents, in other words, are not yet a meaningful source of that growth.

Why so few people are using agents

Uber offered one practical explanation: its app is already quick. The company said the average time to book a ride in the app is about 50 seconds, and reordering a favourite meal takes around 90 seconds. When the regular path is that short, there is little time for an AI agent to save.

Chief executive Dara Khosrowshahi was candid about the early results. He described the traction as very early and very limited, and said Uber hopes usage will grow.

Uber is not alone. DoorDash said on its own earnings call the week before that it had received only a few hundred orders from AI agents, with most customers still ordering directly in its app. (DoorDash is also in the middle of a much larger move: its $17 billion deal to buy Instacart.)

Uber is still betting on agents

Despite the small numbers, Uber is positioning itself for an agent-driven future rather than resisting it. Khosrowshahi said he sees agents as another way for customers to reach Uber, and expects them to add to demand overall rather than take it away.

He noted that there are currently around eight major AI assistants on the market and predicted the number will grow beyond ten. He also said every major AI company is talking to Uber. According to coverage of the earnings, Uber is building its own AI agent experiences while also working with outside assistants from companies including Apple, OpenAI, Anthropic and Google.

Inside its own apps, Uber is already using AI in other ways, including a shopping cart assistant, destination prediction, and agents that help drivers and couriers decide where and when to work.

Why it matters

The agentic AI boom is built on the idea that people will hand everyday tasks to software that acts on their behalf. Major AI labs, including OpenAI, Anthropic, Google and SpaceXAI, now offer their own agents, and startups in the space have raised money at very high valuations.

Uber’s data is an early reality check. Real-world usage of agents for simple consumer tasks is still close to zero at one of the world’s largest consumer platforms. That does not mean the idea will fail, but it suggests adoption will take longer than the hype implies, especially for tasks that are already fast and familiar in an app.

For businesses, the lesson is to prepare without overinvesting. Making services easy for agents to use is likely to matter over time, but customers are still overwhelmingly coming through traditional channels today.

What happens next

Uber said it expects to keep integrating with AI assistants as they mature. Other consumer platforms, such as Airbnb, are taking a different approach by building their own travel agent instead of relying on general-purpose assistants. The coming quarters will show whether agent-placed orders start to climb or stay a rounding error.