RBNB challenges chatbots, saying travel experience requires collaborative interfaces and an agent operating system

The “RBNB” platform launched an AI-powered search engine as part of its latest update, a step that reflects the platform’s cautious and deliberate stance toward adopting these technologies. Brian Chesky, co-founder and CEO, explained in an interview with the Tech Crunch platform that single-user automated chat interfaces are unsuitable for e-commerce and travel-option exploration, considering that limiting interaction to a text-based dialog window is a short-sighted approach that does not respond to shoppers’ behavior.

Chesky believes that the flaw of chatbots in commerce lies in presenting limited options each time, forcing the buyer to undergo multiple conversation rounds to reach the desired outcome. Travel studies show that the pleasure of planning and browsing options often exceeds the pleasure of the trip itself, so concealing the visual search library removes an element of inspiration and reflection.Current chat interfaces are designed for a single user, while travel decisions rely on family and group participation.Therefore, the company has begun exploring interfaces that support collaborative work among multiple people simultaneously in the coming months.

While startups test generative screens built instantly, Chesky predicts a hybrid that combines predefined interfaces with generative ones. The RBNB CEO rejects the prevailing belief that visual interfaces will disappear in favor of typed commands, asserting that designers remain better able than ordinary users to steer software efficiently, and that major apps retaining their visual identity and distinction is a necessity driven by specific functions such as identity verification, map browsing, and host communication.

The platform is preparing to work with consumer-focused AI agents such as Instank and Muse, treating them as sources for acquiring prospective customers, noting that integrating these tools requires a robust software infrastructure that enables smooth transaction handoff to the app interface. The company is developing specialized agents for customer service and exploratory categorization, culminating in a primary agent that communicates with other application agents via the Model Context Protocol (MCP), the software standard for linking AI agents with external software.

The biggest dilemma lies in the sector’s lack of a genuine operating system built at the kernel level for AI.Chesky sees the race to build the primary agent hampered by the absence of integrated software development kits, citing OpenAI’s former app store experience that lacked a cohesive runtime environment, and emphasizing that the fundamental shift from standalone apps to agents remains dependent on major firms launching new operating platforms, alongside the growing reliance on voice as a primary interaction tool through voice agents deployed by the platform for customer service and search.

This shift requires product leaders and engineers in the Gulf, Egypt, and the Levant to reassess store interfaces and digital booking platforms. Rushing to replace browsing screens with chat windows has proven ineffective for purchase decisions and comparative shopping, highlighting the need for interfaces that support collaborative planning suited to local spending habits and common family travel decisions. The phase also demands that local platforms prepare back-end software architectures to receive requests from independent agents via open integration protocols, without compromising payment, identity controls, or direct user interaction.