First, What Is an AI Agent?
Think of an AI agent as a smart, autonomous assistant. Unlike a simple chatbot that answers questions based on a script, an AI agent can understand a goal, make a plan, use different tools (like accessing a calendar or a booking system), and take action
to complete a task on its own. These agents can handle complex, multi-step processes, from ordering groceries based on a handwritten list to coordinating logistics between thousands of drivers. For companies operating at a massive scale, this ability to automate complex decision-making is a game-changer, promising unprecedented efficiency.
Uber: The Scale Is the Story
During one week around Halloween, Uber processed a staggering 320 million trips and deliveries combined. Yet, the number of those transactions initiated by external, third-party AI agents was surprisingly tiny—just four rides and about 500 Uber Eats orders. This doesn't mean AI isn't critical to Uber; it just highlights a different strategy. The real story isn't about customers using external AI to book rides. It's about how Uber uses its own internal AI agents to manage the immense complexity of its network. Uber has long used AI for matching riders with drivers, optimizing routes, and dynamic pricing. Now, it's deploying more advanced agents to handle internal engineering tasks, automate bug fixes, and even help secure its platform against new AI-related threats. For Uber, AI agents are less of a customer-facing novelty and more of a core operational engine.
DoorDash: From Text to Table
DoorDash is taking a more direct-to-consumer approach with its AI agents. The company recently launched a text-to-order agent that allows users to place orders via Apple Messages without ever opening the DoorDash app. You can text 'order my usual' and the agent will access your order history, build a cart, and use your saved payment info. While CEO Tony Xu noted that adoption is still in its early stages—with only several hundred orders placed this way so far—it signals a clear ambition to reduce friction. Beyond consumer ordering, DoorDash is also using AI to supercharge its support contact centers, using agents to handle hundreds of thousands of 'Dasher' inquiries daily, which speeds up issue resolution and frees up human agents for more complex problems.
Airbnb: The AI Travel Concierge
For Airbnb, the promise of AI agents lies in perfecting the entire travel journey. CEO Brian Chesky has been open about the company's goal to build a comprehensive AI travel agent, with a launch planned for the first half of 2027. The vision is to create an agent that does more than just book a stay; it will help guests plan trips, compare properties, and manage their bookings. The agent will also provide 24/7 support for issues like changing a reservation or getting help with a refund. On the host side, AI assistants are already emerging to automate guest messaging, answer frequently asked questions, and coordinate with cleaning crews, turning the complex logistics of managing a property into a more automated process. While bookings from external AI agents are currently 'very non-material', Airbnb is betting that a deeply integrated, proprietary AI will become an indispensable part of its service.
A Cautious Revolution
The data from Uber and DoorDash shows that consumers aren't yet rushing to use third-party AI agents to order rides or food. One reason is that the apps are already incredibly efficient; booking an Uber takes a median of just 50 seconds. However, the low adoption of external agents belies the deeper, internal revolution. These companies are not just experimenting with AI; they are fundamentally re-architecting their operations around it. They are building their own agents, embedding them into core workflows, and using them to manage logistics, customer service, and even software development at a scale previously unimaginable. The customer may not always see the agent, but they will undoubtedly feel its impact through faster, smarter, and more personalized services.
















