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

Only four Uber rides were ordered through an AI agent during the company's busiest week of October — the week of Halloween — according to figures Uber shared on its third-quarter earnings call on Wednesday.

Four. Out of roughly 320 million trips and orders across Uber and Uber Eats that same week.

The number came from Uber's chief financial officer, Balaji Krishnamurthy, who told Wall Street analysts that the company handled around 320 million trips and orders during the week of Halloween, but only 500 of those were Uber Eats orders placed by an AI agent. The rideside figure was even smaller: four trips booked by an agent across the entire platform during what should have been peak demand.

CEO Dara Khosrowshahi did not try to spin it. "It is very, very early and very limited traction that we are seeing so far," he said on the call. "Listen, we hope that there will be more traction."

The disclosure lands at an awkward moment for the AI industry, because the numbers cut directly against the dominant narrative of the moment.

Why do four rides matter?

For the past year, the AI sector's biggest bet has shifted from chatbots that answer questions to agents that act — software that books your ride, orders your dinner, and completes transactions on your behalf. The industry calls this the agentic web, and nearly every major player is now racing into it.

OpenAI, Anthropic, Google, and SpaceXAI (Grok) are all offering their own AI agents, according to the source, alongside numerous smaller startups. There has been an explosion of personal AI startups, including one of the buzziest, Instinct, founded last year, which raised its most recent funding round at a $10 billion valuation. Meta's Muse became the No. 1 app in Apple's U.S. App Store last month and topped 10 million users, per Meta's own reporting. The company said many of those users were employing it to shop and transact with businesses.

Against that backdrop, Uber's numbers are a reality check. If agents can't gain traction on one of the world's largest transaction platforms, during its busiest week of the year, the agentic web still has meaningful distance left to cover. Early figures from other public companies point the same direction.

What are other companies seeing?

DoorDash, for instance, said last week on its own earnings call that it has received only a few hundred orders from AI agents so far. Most customers are still placing orders through the DoorDash app, the company said.

Uber is seeing the same behavior. More people used an AI agent to order food through Uber Eats than to book rides, but the volume remained small compared to overall trips, the company indicated on the call.

The pattern across both companies is consistent: AI-initiated transactions exist, but at volumes that round to zero against the core business.

Is an AI agent even faster than an app?

Khosrowshahi offered one concrete explanation for the anemic numbers, and it's a simple one: the math doesn't favor the agent.

The average time to book a ride in Uber's app is only 50 seconds, he pointed out on the call. Re-ordering a favorite food through Uber Eats takes only 90 seconds. Those are the tasks AI agents are supposed to automate — and the app already does them fast enough that an agent isn't necessarily saving anyone time.

That framing matters for the broader agent debate. The pitch behind agentic AI is efficiency: hand off tedious tasks to software. But when the tedious task takes under a minute, the value proposition narrows to trust, accuracy, and preference — whether a user believes an agent will get the order right, not whether it will save ten seconds.

For now, users appear to be answering with their thumbs. They keep opening the app.

Is Uber still betting on agents?

Despite the low transaction counts, Khosrowshahi sounded bullish on agents as a category. His issue was timing and traction, not the concept.

"We think that is another opportunity to reach our platform. We think it will be a net positive in terms of demand on the ecosystem," he said on the call.

He also gave a sense of how Uber sees the competitive map. There are around eight major AI assistants on the market right now, Khosrowshahi noted, and he expects that number to grow to more than 10 over time.

And Uber, it seems, is not waiting for the field to settle. "Every single AI system company is talking to us," he added.

That last line is the one to watch. If eight-plus assistant platforms — and eventually more than ten — each need ride-hailing and delivery inventory, Uber positions itself as infrastructure that every agent must plug into. The company doesn't need consumers to love agents; it needs agents to need Uber. On that reading, four Halloween rides aren't a failure so much as a baseline, and the real question is what the number looks like when — or if — the agent platforms themselves find their audience.