By Emmanuel Mounier, Secretary General, Global Travel Tech.
A few weeks ago, I wrote that nobody could predict how AI will reshape travel tech. That is still true. But Google’s announcement on 19 May, presented by the company itself as “a new era for AI Search,” gave us a clear signal. The well-known list of suggested blue links are no longer the heart of the product: the default is now a conversational assistant, with AI Overviews sitting on top of the classic results page. TechCrunch was blunt the same day: Google Search as we knew it is over.
Travel is one of the most search-dependent industries online, and Google is now absorbing the whole discovery journey, from comparison and reviews to maps, prices and itineraries, into a single conversation. That raises some real questions for our sector, and some that regulators cannot afford to ignore either. Is the shift from gate to wall actually a good thing? Who wins, between direct and indirect distribution channels, in a world where the AI assistant decides what the traveller sees? And what does any of this mean for the competition rules that were meant to keep Google in check?
Is this a good thing?
Google has always been a gatekeeper of the open web. Not always perfect, but at least there was a gate: users were routed onwards to airlines, hotels, online travel agencies, review sites and independent publishers, each with a relatively fair chance of being found. AI Search changes that logic. When a traveller asks where to spend a long weekend in October, the assistant synthesises destinations, weather, hotel options, prices and reviews inside its own answer, using content from third party sites. Pew Research Center found that when an AI summary appears, users click a traditional result in only 8% of cases, against 15% when no summary is shown. Reuters Institute expects publishers’ search traffic to fall by more than 40% over the next three years.
If those numbers holds up, the free travel web does not completely disappear, but it becomes harder to reach. The gate starts to look more like a wall, with the rest of the internet quietly tucked behind it.
Who wins between direct and indirect channels?
Bookings still have to happen somewhere. So the real question is who gets surfaced inside the AI assistant before the traveller ever sees a booking page. Google has already announced partnerships with Booking.com, Expedia, Choice, IHG, Marriott and Wyndham, with the stated aim of letting users complete flight and hotel bookings inside AI Mode with the “partner of your choice.” It is also building dedicated ad formats for AI Mode.
If a traveller already knows the airline or hotel chain they want, an AI agent may well route them straight to the supplier. But most leisure queries are exploratory, comparative, uncertain. For that kind of query, AI assistants will probably suggest large online travel agencies and large supplier groups far more often than independent hotels, smaller carriers, or niche tour operators. And because of the conversational format of an AI assistant, this might mean that SMEs in the travel sector will have an even more difficult time to be seen than they already have with the current Google Search.
And let us not forget: this is still a competition problem
It would be tempting to treat all of this as a purely technological story. It is not. Google still holds around 87.5% of the European search market, and a dominant platform redesigning its own interface to keep more of the user journey inside its own walls is exactly the kind of behaviour competition law was built to scrutinise. If anything, the new interface makes the regulators’ job harder, because the conduct shifts from ranking decisions we can observe to generative outputs that are inherently more opaque.
European regulators have not been standing still. The Commission’s preliminary findings of March 2025 already concluded that Alphabet favours its own services in Google Search, with travel and hospitality explicitly in scope. The April 2026 DMA review confirmed that when gatekeepers deploy AI inside their core platform services, existing DMA obligations, including the self-preferencing ban, still apply. The Commission has also proposed forcing Google to share search data with third parties, including AI chatbots, on fair and reasonable terms. The principle to defend is simple: dominance follows the interface, not the format. Self-preferencing is the same problem whether it sits on a classic results page or inside an AI answer.
Our Path Forward
AI in travel is genuinely useful, and our sector should not pretend otherwise. Travellers will increasingly expect conversational planning, real-time recommendations and, eventually, agentic booking. The job of the coming years is to make sure this new layer stays contestable, so that more than one company can build on top of it and the rest of the travel ecosystem is not quietly enclosed in the process. We do not need to refuse AI Search. We need to make sure it stays a gate, with more than one key, and one that travellers and travel businesses can still get through.