For the last year, most of the search conversation has been framed as inevitable. Google moves from blue links to AI summaries, users get faster answers, publishers lose traffic, advertisers adapt, and the old search journey fades into the background. The direction feels obvious: search becomes less about sending people to websites and more about answering inside the platform.

Some people are already calling this “the death of the internet.” That sounds dramatic, but the concern is not completely irrational. The web many of us grew up with was built around search results, links, exploration, comparison, and discovery. You searched, opened several pages, checked sources, compared opinions, went down side paths, and built your own view. AI search compresses much of that into one answer box. It is faster, but it also changes the user’s role from explorer to recipient.

The move itself is natural. Clicks from search have been under pressure for a while, especially for informational queries. AI Overviews accelerate the shift from “send users elsewhere” to “answer inside the platform.” Google’s ad business is not collapsing, but the old click-based structure is clearly being renegotiated. Fewer open-ended search journeys means fewer traditional opportunities for publishers, brands, and comparison sites to participate in discovery.

What many of us did not expect is how quickly a visible “No AI” alternative would become attractive. That is the real signal. The issue is not that people reject AI altogether. Most users will accept AI summaries when they are useful. The issue is that some users do not want AI to become the only default interface for finding information. They still want control, links, source diversity, and the ability to search without the answer being pre-shaped for them.

There is an important lesson here for travel.

AI search, AI trip planning, AI concierges, and AI booking assistants will become normal. They are useful. They reduce friction. They can help travellers move from vague intent to a shortlist faster than traditional search ever could. A traveller who says “I need a family-friendly hotel in Lisbon, near restaurants, with parking and flexible cancellation” should not have to open twenty tabs just to build an initial shortlist.

But not every traveller wants the same interface all the time. Sometimes people want the summary. Sometimes they want the map. Sometimes they want ten hotel tabs open. Sometimes they want to compare reviews manually. Sometimes they simply do not trust the machine enough to let it narrow the world for them.

That matters because travel is not a simple information category. It is high-consideration, emotional, expensive, and full of trade-offs. A hotel choice is not just a factual answer. It depends on location, budget, brand perception, reviews, cancellation rules, loyalty benefits, room type, transport, neighbourhood feel, and personal risk tolerance. AI can help structure that complexity, but it should not remove the user’s ability to inspect it.

Travel companies introducing AI into search and booking should not treat “AI-first” as “AI-only.” A better product principle is simple: give users the AI shortcut, but keep the traditional path visible.

For hotels, OTAs, airlines, destinations, and experience platforms, this matters commercially. AI fatigue is already forming. If the interface feels forced, users may not just ignore the AI layer. They may switch to platforms that give them more control. In travel, trust is part of conversion. A traveller who feels boxed into an opaque recommendation may hesitate, even if the recommendation is technically good.

The better approach is hybrid discovery. Use AI to summarize, filter, explain, and reduce friction. But keep maps, filters, source links, review access, comparison tables, and human support available. Let the traveller decide when to use the assistant and when to go back to manual exploration.

This also applies to AI concierges in hotels and destinations. A guest may appreciate an instant answer to “What can I do nearby tonight?” But if the assistant only pushes a narrow set of recommendations, confidence drops. Guests still need alternatives, visible reasoning, and a path to a person when the request becomes specific or sensitive. The same applies to flights, tours, restaurants, transfers, and attraction bookings.

The bigger takeaway is that adoption is not only about adding AI. It is about giving users a choice in how much AI they want in the journey. In travel, the winning interfaces may not be the ones that automate the most. They may be the ones that know when to step back.