The Dawn of Digital Butlers
Forget the simple chatbots you've argued with about a refund. The new frontier is 'AI agents'—smarter, more autonomous programs designed to act as your digital proxy. Think of an assistant you can give a goal to, like 'order my usual Friday night dinner',
and it will handle the entire process without you needing to open an app, browse menus, or enter payment details. This is the vision companies like Uber, DoorDash, and Airbnb are chasing: an evolution from you tapping through their apps to you simply telling an AI what you want, and the agent making it happen. The core idea is to move beyond responding to clicks and toward understanding and executing complex intentions.
A Few Hundred Orders: A Small Start
The headline-making number comes directly from DoorDash's recent earnings call, where CEO Tony Xu revealed that third-party AI agents had driven 'maybe several hundred orders' in total. In the context of a company that processes nearly a billion orders per quarter, this figure is statistically tiny. Similarly, Uber reported that during its busy Halloween week, AI agents booked only four rides and about 500 Uber Eats orders out of 320 million total transactions. While these numbers seem almost comically low, they represent the first real-world data points for a technology that has been mostly theoretical. Executives are framing this as a 'very low use case' for now, acknowledging that consumer habits are hard to change when ordering a meal already takes just 90 seconds in an app.
Not Just Outsiders: Building In-House AI
While the use of external agents like Meta's Muse or others is minimal, it hasn't stopped these platforms from aggressively developing their own AI capabilities. DoorDash, for instance, is testing a 'Text DoorDash' feature that allows ordering via text message and has an in-app 'Ask DoorDash' assistant. Uber Eats has a 'Cart Assistance' feature where AI fills your basket from a photo of a recipe. These proprietary tools are seeing more success, with Uber noting that baskets created by its AI are often twice the value of average orders. The strategy appears to be two-pronged: allow limited access to outside agents to understand the trend, but build a superior, integrated experience to keep customers within their own ecosystem.
Airbnb's Vision: The AI Travel Planner
Airbnb is taking a different, more ambitious approach. CEO Brian Chesky has been dismissive of 'non-material' bookings from external agents, arguing that travel planning is a fun, visual, and social process that generic, text-based agents ruin. In response, Airbnb is building its own AI-native travel agent, slated for a 2027 launch. Instead of a chat window, this agent will generate a dynamic user interface in real time, pulling together maps, photos, calendars, and price comparisons based on your conversation. The goal is to handle everything from initial inspiration to post-booking support like changes and refunds, acting as a full-service concierge that preserves the joy of planning a trip.
The Real Battle: Who Owns the Customer?
The cautious adoption rates and simultaneous in-house development reveal the central tension of the AI agent era: who will own the customer relationship? If users begin relying on a single, third-party AI assistant to order rides, food, and stays, then Uber, DoorDash, and Airbnb risk becoming interchangeable 'fulfillment layers'. Their brands, loyalty programs, and direct relationships with consumers could erode. By setting the rules for agent access and building their own compelling AI experiences, these platforms are positioning themselves to control the future. The few hundred orders placed today are less a measure of current demand and more the opening moves in a long-term strategic contest for the front door to digital commerce.
















