During Halloween week, third-party AI agents booked four rides on Uber, according to CFO Balaji Krishnamurthy. That same week, Uber and Uber Eats handled roughly 320 million combined trips and orders.
The agent-booked Uber Eats count reached approximately 500 orders, but the contrast with total platform volume remains stark. CFO Krishnamurthy described the overall result as “very, very early and very limited traction,” while indicating the company expects adoption to grow.
DoorDash and Airbnb Report Similar Results
Across three major platforms, agents are live, but consumers are largely not using them.
DoorDash CEO Tony Xu told analysts that third-party AI agents had generated “maybe several hundred orders in total” since becoming available, measured against the billions of orders DoorDash processes globally each year.
Airbnb CEO Brian Chesky said only a “very non-material number” of the company’s nearly 500,000 daily bookings came through AI agents. All three disclosures cover the earliest window of public availability: Meta’s Muse launched September 8 and OpenAI’s Dots launched September 29, according to TechSpot.
Why Consumers Are Not Delegating
Speed and personal preference are working against the agents, at least for now.
Uber executives offered a practical explanation, noting that routine transactions are already fast. That framing weakens the convenience argument for handing those same tasks to an external agent.
Chesky identified a different obstacle for Airbnb. Bookings typically involve personal research into neighborhoods, property photos, and price, and many users prefer to make those calls themselves rather than delegate them to automation.
A third factor is trust. Delegating a purchase requires confidence that an agent will select correctly, respect constraints, and stop before completing an unwanted transaction.
What the Data Actually Establish
Negligible early adoption is not proof of permanent failure, but it is proof that behavior has not changed yet.
The figures reflect a launch period, not a mature market. Meta’s Muse and OpenAI’s Dots had been publicly available for only weeks when executives reported these numbers.
Uber, DoorDash, and Airbnb are each developing their own AI tools, according to Business Insider. Platforms that embed AI within existing accounts, payment systems, and loyalty programs may hold a structural advantage. External agents must ask users to extend trust to an entirely new intermediary.
Where Things Stand
Capable technology, familiar apps, and not enough reason yet to change.
AI agents can technically book a ride, place a food order, or reserve an accommodation. What the transaction data show is that consumers have not yet found a compelling reason to stop using apps that already complete those tasks quickly.
Four rides and several hundred food orders against hundreds of millions of weekly transactions is not a rounding error. The gap between the industry’s ambitions for agentic commerce and the figures disclosed by three of its most prominent platform partners is, at this moment, very wide.
