Defining the 'AI Agent'
Before diving into the numbers, it's crucial to understand what an 'AI agent' is in this context. It is not just a chatbot that answers questions. An AI agent is a piece of software designed to act on a user's behalf, autonomously performing tasks like
researching options, making decisions, and completing purchases. Think of a personal assistant that you can give a goal to—like 'order me a vegetarian pizza' or 'book a ride to the airport'—and it handles the entire process without you needing to open an app and click through the steps yourself. These can be third-party agents like Meta's Muse or tools built directly by companies like DoorDash's own 'Ask DoorDash' feature.
A Few Hundred Orders: Hype vs. Reality
During a recent earnings call, DoorDash CEO Tony Xu revealed that third-party AI agents have been responsible for 'maybe several hundred orders in total' since they have been active. On its face, that might sound like a significant step into the future. However, the reality requires context. DoorDash processes a colossal volume of transactions—in 2025, it handled roughly 3 million deliveries per day, and in the third quarter of 2026 alone, it processed nearly one billion orders. In light of these figures, 'a few hundred' cumulative orders is a statistical drop in the ocean. It's clear that this technology is not yet a meaningful driver of business for DoorDash. CEO Tony Xu himself called it a 'very low use case' for now, noting that customers overwhelmingly still use the main app for routine ordering.
Uber and Airbnb Tell a Similar Story
DoorDash is not alone in these findings. Other platform giants are seeing similarly low adoption rates for third-party AI agents. During Halloween week, one of its busiest periods, Uber reported only four rides and about 500 Uber Eats orders were placed by AI agents out of a staggering 320 million total trips and deliveries. Likewise, Airbnb CEO Brian Chesky described the number of bookings made through agents as 'very non-material' compared to the platform's nearly 500,000 daily bookings. The consensus among these executives is that for simple, fast tasks like ordering a car or a meal—which can take less than 90 seconds in an app—users don't yet see a compelling reason to delegate the task to an AI.
Why Bother With AI Agents At All?
If the numbers are so small, why is this even a conversation? Because it's about the future, not the present. While external agents are seeing little traction, these companies are heavily investing in their own AI capabilities. DoorDash has launched 'Ask DoorDash' and a text-to-order service, which let users place orders using natural language. Similarly, Airbnb is building its own specialized AI travel agent, set to roll out in 2027, believing that the best agents will be specialists built by companies that know their users. By opening their platforms in a controlled way to third-party agents, these companies are gathering data, testing the technology, and positioning themselves to control the 'front door' to their services, no matter what interface customers use in the future.
The Road From Hundreds to Millions
The path to widespread adoption of AI agents for commerce is still long. The current low numbers reflect that consumers haven't yet built the trust or habit of using these tools for purchases. Furthermore, some tasks, like planning a vacation, are activities that users enjoy and may not want to fully automate. For AI agents to move from being a curiosity to a mainstream channel, they must prove they are not just capable, but also more convenient, reliable, and trustworthy than the highly optimized apps users are already accustomed to. Companies like DoorDash, Uber, and Airbnb are betting that by engaging now, they can shape how this technology evolves and ensure they remain central to the transaction, even if the customer's journey starts with a conversation with an AI.
















