Uber says third-party AI agents booked just four rides last quarter

Four rides in a quarter of 320 million

Uber has, for the first time, put a number on how much of its business arrives through third-party AI agents, and the figure is strikingly small. On its third-quarter 2026 earnings call, executives said autonomous agents booked just four rides over the entire quarter, according to The Verge. Agent-placed Uber Eats orders were somewhat more common, at around 500 during the week of Halloween, but that still compares with a total of 320 million combined rides and deliveries for the period. Even under the most generous reading, agent-booked rides accounted for no more than roughly one in 80 million transactions.

The disintermediation worry

The reason anyone is counting at all comes down to a scenario known as disintermediation: consumers ask an AI assistant to call a car or order dinner rather than opening the app itself, weakening the platform's direct relationship with its customers. With the arrival of general-purpose agents such as Meta's Muse and OpenAI's Dots, some experts have pointed at on-demand platforms like Uber and DoorDash as especially exposed, since booking and ordering are among the most obvious things to delegate to an agent. Until now, that argument has rested largely on speculation. Uber's disclosure offers one of the first concrete baselines from a major consumer platform, and it suggests the exposure is, at this stage, theoretical.

Uber's position: open to partners, not worried

Executives said Uber is willing to work with third-party AI companies to make its platform easier for agents to use, while expressing confidence that customers will keep preferring the company's own app. "We hope that there will be more traction," CEO Dara Khosrowshahi said on the call, arguing that Uber's direct services will remain the smoothest way to use the service but that the company "will work with these third parties." He described the moment as "very, very early" and said Uber wants to avoid becoming dependent on any single player. As The Verge observes, platform operators also have little incentive to let agents barge in and replace their own customers, and today's first wave of agents still carries enough friction that most people simply default to the app.

Uber is building its own AI anyway

Rather than waiting to find out, Uber is hedging with in-house AI designed to keep users inside its own surfaces. Earlier this year the company rolled out a Cart Assistant for Uber Eats that turns a photo of a recipe or a text prompt into a grocery list. The subtext is clear: if agent-mediated ordering ever does take off, Uber intends to own that layer rather than cede it to an outside assistant.

Why it matters

Hard adoption numbers for AI agents are almost never published. Vendors describe agentic commerce as inevitable while revealing very little about actual transaction volumes, so Uber's four rides and few hundred orders set against 320 million trips is the clearest datapoint yet that, in late 2026, consumer agent adoption remains marginal even in the use case widely considered the most natural fit. The disclosure also shows where the leverage sits: agents need cooperation from platforms, in the form of APIs, partnerships and terms that Uber can grant or withhold, while Uber simultaneously builds competing features of its own. If Uber keeps reporting this figure, it becomes a genuine benchmark for whether agentic commerce turns into a real shift in how people book rides and order food, or remains a roadmap talking point, and whether on-demand platforms get disintermediated or simply gain one more booking channel they control.