Brian Chesky says Airbnb didn’t integrate its app with ChatGPT because ChatGPT integrations aren’t “quite ready” yet, and he advised Sam Altman on ChatGPT apps
Airbnb CEO Brian Chesky is tapping the brakes on deep, in-app integrations with ChatGPT. In recent remarks, he said the company has not wired Airbnb’s core experience into ChatGPT because the current generation of integrations isn’t “quite ready” for mass-market use. Chesky also said he’s shared product advice with OpenAI CEO Sam Altman about how “ChatGPT apps” could work best for consumers and developers.
Why Airbnb is waiting
Chesky’s stance reflects a pragmatic read of where generative AI stands: impressive, fast-evolving, and increasingly multimodal—but still maturing on reliability, latency, predictability, and privacy controls at the scale a platform like Airbnb demands. For a service that handles high-stakes transactions, identity verification, and real-world travel logistics, even occasional hallucinations, incomplete context hand-offs, or API instability can erode trust.
There are also user-experience trade-offs. If a ChatGPT-powered flow adds steps, introduces lag, or creates uncertainty about data use, it can degrade the polished, guided paths Airbnb has refined over years. Chesky’s calculus appears to be that AI should feel like a clear upgrade—faster, more accurate, and more personal—before it sits at the center of the product.
What he told Sam Altman
Chesky said he offered Altman feedback on “ChatGPT apps”—the emerging idea that third parties can build lightweight, specialized experiences that live inside ChatGPT. While he didn’t divulge specifics, the throughline is easy to infer from Chesky’s product philosophy: keep the experience useful and intuitive, make hand-offs seamless, ensure data handling is transparent, and prioritize reliability over novelty. For consumer apps, that typically means nailing a few end-to-end jobs (plan a trip, resolve an issue, change a booking) instead of scattering shallow features across many use cases.
The state of ChatGPT integrations
The industry has already cycled through a few approaches—plugins, custom GPTs, and evolving developer APIs. Each wave makes it easier to connect proprietary data, automate routines, and surface structured actions. But the final 10%—the difference between a dazzling demo and a dependable, day-in, day-out assistant—remains the hardest. Companies with complex transactional systems need:
- Consistent reasoning and up-to-date context across long sessions
- Low-latency responses and graceful fallbacks when AI is unsure
- Privacy, auditability, and granular consent for sensitive data
- Strong guardrails to prevent mistaken actions or confusing advice
Until those pieces are nailed, many consumer platforms are experimenting at the edges—AI for support, smarter search and recommendations, content summarization—rather than plugging AI into the heart of mission-critical flows.
What an ideal Airbnb–ChatGPT experience could look like
The opportunity remains huge. A great integration could combine rich travel data, user preferences, and real-time availability to:
- Co-plan itineraries with live pricing, host responsiveness, and neighborhood insights
- Bundle stays with transportation and experiences, adjusting plan options in natural language
- Handle post-booking changes—date shifts, guest count updates, special requests—without friction
- Proactively flag policy nuances, fees, and house rules to prevent surprises
But for that to work, the assistant needs reliable action-taking, clear provenance for claims, and a smooth hand-off between conversational guidance and transactional steps.
Why Chesky’s caution matters
Airbnb is a bellwether for consumer marketplaces. If it waits, others may pause full-stack integrations too, focusing instead on safe, high-ROI use cases behind the scenes (fraud detection, content moderation, listing quality, and support tooling). That doesn’t signal skepticism about AI’s potential—only that trust-heavy experiences demand a higher bar for “ready.”
What to watch next
- Reliability and traceability: Can assistants cite sources, confirm actions, and gracefully decline when unsure?
- Performance: Sub-second responses for common flows and stable APIs under real-world load
- Privacy and compliance: Clear data boundaries, enterprise controls, and user-consent UX
- Design patterns: Native-feeling hand-offs between chat and app screens without context loss
When those foundations mature, expect companies like Airbnb to move quickly. Conversational interfaces are well-suited to the ambiguity of travel planning; the question isn’t whether Airbnb will integrate deeply, but when the experience becomes undeniably better than today’s touch-first flows.
Chesky’s message lands somewhere between optimism and restraint: AI assistants will transform consumer apps, but the best integrations won’t be rushed. For now, Airbnb will keep experimenting—just not at the expense of the clarity and trust that make people comfortable booking a home halfway around the world.
