Reading time: 4 minutes
Don't be fooled. What OpenAI has created is essentially your Booking mobile app, but inside a ChatGPT window. They are making the classic mistake of digital transformation: "digitize" existing processes instead of completely reimagining the experience, wasting the disruptive moment we are experiencing.
This is not genuine progress or disruptive innovation. ChatGPT, through the Booking and Expedia apps, is not reimagining the booking experience. You are “hosting” outdated existing reserve interfaces and processes.
The immediate future is not limited to that, but points to truly immersive experiences: visualizations, image and video in real time, augmented reality, simulations of the destination before booking, etc. But as long as that almost science fiction scenario of the Customer Journey arrives, we are still trapped in the old paradigm, no matter how big the greats want to decorate it.
The silent war for distribution
We are here, at this point. Language models (LLMS) do not connect directly with individual hotels.
The Otas, the perpetual problem
At the moment, it seems that only OTAs have privileged access to LLMS. Through "patches" like these apps embedded within ChatGPT, Booking and Expedia are presented as the only ones? Valid distribution channels in the conversational era. But let's analyze what is really happening: this integration is extremely limited. It works only within ChatGPT, leaving out Claude, Gemini, Perplexity, Grok and the rest of the LLM ecosystem. Worse yet, forces the user to manually install an app before using it and only inside chatGPT. It's absurd. This is not natural or technological evolution.
It is a mirage. A commercial move designed to launch an integration marketplace, exactly like the CustomGPT, which never took off.
What we see today is not the future of hotel distribution, it is an advertising product while OpenAI seeks its monopoly.
Google, the silent giant
Google has been quietly building the perfect infrastructure for this moment. Through Google My Business and Google Hotel Ads, it already has a direct relationship with millions of hotel establishments around the world and functions as a hub of connections between hotels and their distribution systems. Today, hotels are already connected directly with Google to serve as a source of data and with a layer of reserve engines such as Neobookings, which allow the reservation to be completed.
The logical leap would be for Google to develop a universal MCP, becoming the standard interoperability layer between hotels and LLMS. All this makes more sense, precisely on the day that it finishes deploying throughout the world (including Spain) its "AI mode", and with it has completely transformed the search experience as we knew it. Powered by Gemini, this new mode abandons the classic list of blue links to offer something radically different: direct, detailed and conversational answers that understand the context and allow becoming one more covert LLM.
The aggregators, a new actor who enters the scene
On the other hand, the aggregator platforms also want to gain a foothold and are trying to be one more player, promising direct connections without such abusive commissions. But let's be honest: they are competing against giants who invest billions in technology and marketing. By the time the aggregators have agreed on who will be the leader (we return to the fight separately), it will be too late again.
Hotel + CRS integrated with MCP technology
the desired solution It happens by democratizing your hotel's access to artificial intelligence, and I am convinced that it will be so. Imagine that each hotel can expose itself to any LLM and offer all its information, ARI and the possibility of ending the reservation directly. This is as it happens today with any website of a hotel, it is indexable by search engines and they have a reservation engine. No intermediaries. No excessive fees.
Imagine the following flow: AI, instead of limiting itself to information from OTAs, Google and aggregators, performs an additional verification: Chatbots (Claude, Gemini, ChatGPT) will eventually learn how to ping the hotel's web servers, Detect if there is an MCP server behind them, and deploy it in real time as an additional data source to the conversation, allowing direct reservation.
-
My vision is that there will be four possible scenarios (in the next 18 months)
- Total OTAs domain (~10% probability)
The OTAs consolidate their dominant position in the conversational space, further increasing their control over hotel distribution and perpetuating dependency. - Google takes control via agentic (~40% chance)
The tech giant leverages its existing infrastructure to connect hotels directly with their AI systems. Another movement we saw coming. - New aggregators fragment the market (~10% chance)
Alternative platforms emerge that promise lower commissions, but establish new dependencies and open a new unnecessary battle. - Direct Window: Hotels with MCP eliminate intermediaries (~40% probability)
The establishments, through their technological partners such as neobookings, implement MCP protocols that allow direct connection with LLMS.
- Total OTAs domain (~10% probability)
At the time we wrote this article, Bloomberg published a report with the first impressions of use of the Expedia and Booking.com apps within ChatGPT:
- poor experience: Frequent errors and rigid commands make reservations less intuitive than on traditional sites.
- Limited functionality: Only interact with one app at a time, making real-time comparisons impossible.
- Data control risk: Hotels lose visibility on how their inventory is displayed and what drives conversions.
- distrust of partners: Brands see it as a complementary channel, not as a replacement for their apps or Google, due to their technical failures and low conversion.
Conclusion: This is a strategic positioning
Although MCP is still maturing and its evolution is uncertain, having MCP-capable systems strategically positions you for when it becomes the industry standard. The question is not whether this change will come, but when. Hotels that anticipate this transition and prepare their technological infrastructure will now be positioned to capture direct demand in an environment where transactional friction is reduced to its minimum expression.
Don't wait for the Otas to decide your future for you once again. If you want to take control now, we are here to show you how.
Raúl Blanco Díaz is Chief Artificial Intelligence Officer at NeoBookings. In its day to day, it investigates, tests and provides vision on the application of artificial intelligence to the hotel sector.