ChatGPT's connected travel apps often go unused even when authenticated, while Claude consistently surfaces and uses them with minimal friction. A hands-on test by Skift found that availability inside an AI chatbot does not guarantee visibility or actual use, creating a significant gap between brands that have integrated with these platforms and those that actually get recommended to travelers.

Why Are Travel Apps Failing Inside ChatGPT?

Skift tested how travel integrations from Booking.com, Expedia, and Viator actually perform inside ChatGPT and Claude, and the results revealed a troubling pattern. Even when these apps were fully connected and authenticated within ChatGPT, the chatbot initially denied having access to them, offering elaborate but false technical excuses. Users had to push back forcefully, in some cases typing "You are wrong," before the apps would actually function.

Expedia confirmed that this behavior is a known bug on OpenAI's platform. The issue highlights a critical vulnerability in the emerging AI-powered travel booking ecosystem: being integrated into a chatbot is not the same as being discovered or used by one. For travel brands, this creates a new layer of referral risk that sits alongside traditional search engine optimization and app store visibility.

Claude, by contrast, handled the same integrations with significantly better performance. The Anthropic chatbot proactively suggested relevant connectors in context and let users activate and use them with minimal friction. This performance gap suggests that the underlying architecture and decision-making logic of each platform shapes which brands travelers actually reach first.

How Are Travel Brands Adapting to AI Chatbot Integration?

• Direct App Connection: Major travel companies including Booking.com, Expedia, and Viator have rushed to add apps to ChatGPT and Claude, enabling live listings, real pricing, and deep-linked booking buttons directly within the chatbot interface.
• Connector Optimization: Brands must now understand not just where their integrations appear, but why and how often the chatbot chooses to surface them, since availability does not equal visibility.
• Platform-Specific Strategy: Travel companies need to develop separate strategies for each platform, as OpenAI and Anthropic have fundamentally different commercial models and technical implementations.

What Does the Commercial Model Look Like?

The commercial landscape for AI chatbot integrations remains unsettled and differs significantly between platforms. OpenAI is layering in advertising, with Booking.com and Expedia ads appearing under rival results within ChatGPT. Anthropic, by contrast, has pledged to remain ad-free, meaning brands cannot rely on paid placement to boost visibility.

This divergence creates a strategic dilemma for travel brands. On OpenAI's platform, brands can potentially pay for visibility, but they must also contend with the platform bug that causes ChatGPT to deny access to connected apps. On Anthropic's platform, there is no advertising option, but the chatbot is more likely to actually use the integrations when they are relevant.

The stakes are significant because chatbot answers increasingly shape which brands travelers reach first. As AI chatbots become a primary discovery channel for travel bookings, the ability to be surfaced and used by these systems has become as important as traditional search rankings. Brands that are connected but not surfaced are essentially invisible, even though they have invested in integration.

What Should Travel Brands Know About AI Chatbot Visibility?

The Skift testing revealed that the current state of AI chatbot integrations is fragile and inconsistent. Brands cannot assume that connecting their app to ChatGPT or Claude will result in actual usage. Instead, they must actively monitor whether the chatbot is surfacing their integration, test the user experience across different query types, and be prepared to optimize based on platform-specific behavior.

For travelers, this means that the cheapest or best option may not be the one the chatbot recommends, depending on which platform they are using and how well that platform's algorithm chooses to surface available integrations. The fragmentation of the AI chatbot landscape is creating new winners and losers in travel booking, and the winners are not always the brands with the best products or prices, but rather those whose integrations work most reliably within each platform's unique technical environment.