Key Takeaways

  • Brian Chesky, CEO of Airbnb, makes a "radical statement" that chatbots are the wrong interface for travel and e-commerce, primarily due to their text-based nature.
  • Chatbots fail at core consumer needs: they lack direct manipulation, make comparison shopping difficult, and offer limited visual interaction.
  • Airbnb's own experience shutting down third-party chatbot app integrations last year suggests internal validation of these interface problems.
  • Chesky envisions a future where "agents"—not apps—drive consumer AI, featuring rich, immersive, and highly visual user interfaces that move beyond simple text prompts.

Chatbots Are the Wrong Interface for Almost Everything Else

Brian Chesky isn't just questioning the prevailing AI interface; he's dismissing it outright. For the ambitious founder building a consumer product, his words are a stark warning: current chatbots are a dead end for industries like travel and e-commerce. “I do not think a chatbot is the right interface for travel or e-commerce,” Chesky states, calling it a "radical statement." This isn't theoretical. Airbnb even launched third-party apps with chatbot integrations last year, only to "shut them down." That's real-world data indicating a fundamental flaw in the text-first approach.

Chesky breaks down the problem into four critical areas. First, chatbots are inherently text-based. In a world saturated with rich media, “photos are an afterthought,” he explains. Imagine trying to book a unique Airbnb stay or browse a complex product catalog using only text prompts. It quickly becomes cumbersome and visually sterile, failing to convey the essential details that drive consumer choices.

Second, there's no direct manipulation. Forget filters, clicking through options, or rearranging elements on a screen. Every interaction demands a new typed prompt. “You have to type every single prompt, which is fine for a conversation, but you can't like add filters. You can't click around,” Chesky says. This friction kills usability for tasks that require quick adjustments, visual feedback, and intuitive control, a staple of modern digital experiences.

Third, comparison shopping is nearly impossible. Most e-commerce and travel decisions involve weighing dozens of options. How do you stack up competing listings or products side-by-side in a text-based chat? “A lot of e-commerce and travel is comparison shopping,” Chesky notes, highlighting how chatbots inherently fail at this fundamental user need to evaluate and choose from a set of alternatives.

Finally, most chatbots are single-player experiences, isolated from broader context or shared decision-making. These limitations mean that while chatbots might excel at direct Q&A, they fall short for complex, multi-modal, or collaborative consumer tasks.

The Future Is Agents, Not Apps

So, if not chatbots, then what? Chesky sketches a future where "agents" replace apps. He insists these won't be "text forward." Instead, they'll feature "really rich user interface[s]." Think beyond conversational AI to generative AI that crafts entire visual experiences in real-time. With advancements in image and video generation models, these agents could offer immersive, interactive environments where users manipulate visual elements directly, compare options intuitively, and engage with AI in ways far more natural than typing commands.

Chesky observes that “almost every AI company is an enterprise company.” This focus, he argues, has left a massive gap in consumer-facing AI experiences, which largely haven't moved "beyond a chatbot." He believes the “consumer is a massive opportunity for AI,” but only if builders embrace a "richer user interface" that moves past current limitations. This vision suggests a world where AI doesn't just respond to requests but actively curates and presents complex visual information, dynamically adapting to user intent without endless typing.

What to Do With This

Stop prototyping your next AI feature as a chat window. This week, pick a core consumer problem you're trying to solve. Now, sketch out an AI interface for it where the user never types a single word, instead relying purely on visual cues, direct manipulation, and generated imagery or video. Force yourself to design an "agent" experience that prioritizes rich, visual UI over text, asking how generative AI can enable intuitive control and comparison shopping within your product.