The Hype Versus the Reality
The idea of a personal AI assistant that can book rides, order food, and plan trips by simply understanding your request has been a long-held dream in tech. These 'AI agents' are designed to be proactive problem-solvers, working across different services
without you needing to open a single app. However, recent earnings calls from the biggest names in the app economy paint a surprisingly different picture. Uber, for example, revealed that during its busy Halloween week, third-party AI agents were used to book a grand total of just four rides. For context, the platform handled 320 million trips and deliveries in that same period. It’s a statistic so small, it’s almost comical.
A Similar Story Across the Board
This isn't just an Uber phenomenon. DoorDash CEO Tony Xu noted that since AI agents became available, they've generated only 'several hundred orders in total'. Similarly, Airbnb CEO Brian Chesky described the number of bookings made through these agents as 'very non-material' compared to the platform's nearly half a million daily bookings. While these companies are all investing heavily in their own internal AI tools to enhance their services, the idea of an outside agent replacing the app interface has, so far, failed to launch.
The App Is Already Fast and Familiar
One of the primary reasons for the low adoption is surprisingly simple: the current apps are already incredibly efficient. Uber CEO Dara Khosrowshahi pointed out that it takes a median of 50 seconds to book a ride and 90 seconds to reorder a favorite meal. For many users, navigating a well-designed app is faster and more intuitive than trying to articulate a complex request to a chatbot, which might misunderstand the query. Why delegate a task that's already quick and easy to do yourself? For many, AI agents feel like a solution in search of a problem that doesn't exist.
The Psychology of Control and Trust
There's also a significant psychological barrier. When real money and real-world logistics are involved, people want to feel in control. Booking a trip on Airbnb isn't just a transaction; for many, it's a process of discovery. Users enjoy browsing photos, reading reviews, and comparing locations—tasks an agent might automate, but which are part of the experience. Furthermore, there's the issue of trust. Handing over credit card information and personal details to a third-party agent requires a leap of faith many consumers aren't yet willing to take. Studies show that a vast majority of consumers still prefer human interaction or direct control over AI, fearing that AI is used to save companies money rather than improve service.
It's About Augmentation, Not Replacement
This doesn't mean AI is useless for these companies. On the contrary, they are using it extensively to augment their existing platforms. DoorDash has AI tools that help merchants create their online presence and allows customers to order by text. Uber uses AI to optimize pricing, predict demand, and power customer service bots. And Airbnb is building its own sophisticated AI agent, not as a simple chatbot, but as a rich, visual trip-planning tool integrated directly into its app, set to launch next year. The goal for these companies isn't to be replaced by external agents, but to use AI to make their own apps smarter, faster, and more indispensable.
















