What Are AI Agents, Exactly?
Before diving into the specifics, it's helpful to understand what an 'AI agent' is in this context. Think of it as a smart assistant that can perform tasks on your behalf. Instead of you opening an app, tapping through menus, and confirming a purchase,
you could simply ask an AI agent to do it for you through a conversation. These agents can be built by the companies themselves, like a chatbot within their app, or by external tech companies. The ultimate promise is to make everything from ordering dinner to booking a vacation as simple as sending a text message. These platforms are designed to understand your intent, access information, and execute actions, effectively becoming a new way for customers to interact with services.
Uber's Reality Check: 500 Orders and 4 Rides
The headline's specific figure comes directly from Uber's recent earnings call. During Halloween week, one of its busiest periods, third-party AI agents were responsible for booking just four Uber rides and placing 500 Uber Eats orders. When compared to the roughly 320 million total trips and deliveries handled by the company during that same week, this number is a tiny fraction—less than 0.0002%. Uber CEO Dara Khosrowshahi acknowledged the 'very limited traction' but remains open to AI agents as a potential new source of customers. One reason for the slow adoption might be the efficiency of Uber's own app; Khosrowshahi noted that booking a ride already takes an average of just 50 seconds, making an AI intermediary less of a time-saver than one might think. The company is also developing its own in-house AI, like a 'Cart Assistant' that helps build grocery baskets, which has proven to create larger order sizes.
DoorDash and Airbnb Tread a Similar Path
Uber isn't alone in seeing slow initial uptake. DoorDash CEO Tony Xu reported that third-party AI agents have only generated 'several hundred orders' in total since they began operating on the platform. Like Uber, DoorDash is focusing on building its own AI tools. It recently launched a text-to-order AI agent that learns a user's preferences, finds deals, and can even help coordinate group orders within a message thread. Meanwhile, Airbnb is taking an even more ambitious, in-house approach. CEO Brian Chesky has stated that while outside AI agents bring in a 'very nonmaterial' number of bookings, the company is building its own specialized travel agent. This tool, expected to roll out in the first half of 2027, won't be a simple chatbot. Instead, it will create custom visual interfaces with maps, photos, and calendars to help users plan complex trips, believing travel planning is an experience that general-purpose AI can't replicate.
Why the Cautious Approach?
The slow adoption figures from Uber and DoorDash reveal a strategic deliberation, not a failure of technology. These companies have spent years and billions of dollars building powerful, user-friendly apps that keep customers inside their ecosystem. Handing over the customer relationship to an external AI agent risks losing that direct connection. As Khosrowshahi noted from his experience in the travel industry's reliance on Google, becoming dependent on a third-party gateway can be a dangerous position. Furthermore, these platforms are also actively investing in their own AI for internal and customer-facing functions. Airbnb's AI now resolves roughly 40% of customer support issues without human intervention. Uber has also been cutting its customer service workforce while embracing AI for both 'zero touch' resolutions and 'one touch' support where AI assists human agents. This strategy allows them to increase efficiency while maintaining control over the user experience.
















