Hardly anyone is using AI agents for bookings, according to Uber and Airbnb
Hardly anyone is using AI agents, several companies say, as users question the bots’ utility.
Recent reports from Uber, DoorDash, and Airbnb suggest that AI agents are not being widely adopted for facilitating bookings. These firms disclosed that their AI agents have only processed a tiny fraction of transactions, while millions of bookings are made by actual users. Uber, for instance, logged 320 million rides booked by humans in a week, while its AI agents managed just four.
DoorDash CEO Tony Xu reported that AI agents accounted for several hundred orders, although the platform processes billions of orders annually. Airbnb CEO Brian Chesky noted that only a minor portion of bookings on their platform utilize AI agents, possibly due to users' preference for researching trips themselves. The backlash against AI agents may stem from the perception that the technology is a solution searching for a problem.
Uber CFO Balaji Krishnamurthy pointed out that booking a ride on the platform takes only 50 seconds, indicating that AI agents may not significantly simplify the process. Furthermore, some individuals find the idea of using AI agents counterproductive, arguing that the enjoyment of planning trips or dining out is part of the experience.
Recent studies have shown that AI tools can increase work hours and require more time for correcting AI-generated content, potentially negating any time-saving benefits. Major companies and governments are also showing skepticism towards AI agents, with United Airlines banning them from booking flights, Alibaba halting the use of AI bots like Claude Code, and the Swedish government blocking access to Claude and ChatGPT due to security concerns.
As such, it appears that AI agents are not yet a widespread solution for streamlining mundane tasks, and a significant shift towards their adoption seems to be a long way off.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.