AI Search Has No Page Two: How Large Language Models Decide Which Tour Companies to Recommend
This is part eighteen of our Putting AI to Work in Your Tour Business series. Watch the full series here and read part one here.
For 20 plus years, the game of online visibility came down to one thing: Rank high enough on a search results page and win the click. That game still matters, but a second one has opened up alongside it, and it works nothing like the first.
When a traveler asks an AI assistant for a tour recommendation, the assistant doesn’t hand back ten blue links. It reads a stack of sources, picks a handful of businesses, and names them directly inside its answer. You’re either in that answer, or you’re not part of the conversation at all.
The industry has started calling this generative engine optimization, or GEO.
“SEO was all about getting clicks. This is about getting quoted.”
The shift matters because a traveler reading an AI answer rarely scrolls further. There’s no page two to fall back on.
How AI Models Pick a Business to Name
Four sources feed into that decision, and none of them work exactly like the search rankings you’re used to watching.
The first is training. Large language models absorbed a huge amount of text, articles, reviews, and mentions of businesses from across the web, some of it months old and some of it years old. That static layer sits underneath every answer the model gives.
The second is live search. Most AI tools, Gemini’s AI overview at the top of Google included, run an actual web search before responding, and they favor pages written with clear, direct answers rather than pages built for style.
The third is structured data, the machine-readable labels on your website, like schema markup, that describe your tours, prices, and reviews in a format bots can parse even when a human visitor never sees it.
The fourth is your presence everywhere else: your Google Business Profile, TripAdvisor, the online travel agents you list with, press mentions, and the links that point back to your site. The assistant cross-checks all of it to decide whether you’re real and well reviewed enough to recommend.
No single fix controls all four. But the recommendations below touch at least one or two of them.
Why Your FAQ Might Be Invisible to the Crawlers Reading It
Most tour websites already have an FAQ. The problem is how that FAQ gets built. A lot of accordion-style FAQs load their answers with JavaScript, which means a visitor sees the answer after clicking, but the AI crawler reading your page source never sees it at all. It only sees the question.
There’s a 10-second way to check this yourself.
- Open the page with your FAQ, right-click, and choose view page source.
- Search that page for a specific phrase from one of your answers, not the question, the answer.
- If it shows up, the crawler can read it too.
- If you can only find the question, your answers are hidden behind JavaScript and effectively invisible to AI search.
What This Looks Like When a Tour Company Does It Well
A travel operator worked with a marketing agency that specializes in this kind of optimization, restructuring their on-page content and the data behind it. The reported results: ChatGPT sessions mentioning the company rose nearly 20 percent from their starting point, and traffic to their booking page from AI tools jumped by 3,000 percent. Nearly all of that traffic arrived from travelers already mid-conversation with an AI assistant, landing straight on the booking page, not from a Google search at all.
Another operator, a multi-day tour company, built a long, specific FAQ answering the exact questions travelers tend to ask their AI assistants, things like how much time a route takes or what a deposit requires upfront. They also published a plain markdown page written specifically for AI crawlers, laying out the foundational facts about their business in a format no human visitor would ever need, but every crawler can read cleanly.
The Best Practices Worth Doing Regardless of What AI Does Next
A few things are worth doing no matter how this space evolves.
Keep your business name, location, and description identical across your website, social profiles, and every listing site you appear on. Machines cross-check that consistency, and contradictions give them a reason to skip you for a source they can verify more easily.
Add schema markup, the behind-the-scenes labeling that tells a machine which number on your page is a price and which is a review count. You don’t need to write that code by hand. Ask an AI tool to draft it and hand it to a developer to implement.
And treat your online travel agent listings as more than a distribution channel. They also function as evidence that you’re a real, active business, which matters to a model deciding who to trust.
This work holds up even if AI search changes shape again next year, because every one of these fixes is also good marketing on its own terms. Clear, direct answers help the humans reading your site as much as the machines.
“You are optimizing for being findable and credible, which never goes out of style.”
Treat this as practical, not alarming. The fixes here overlap heavily with plain good marketing: Clear answers, consistent information, and a website that’s readable by whatever’s reading it, human or machine.
Three Things to Do This Week
Run a visibility test. Ask two or three AI tools the same five or six questions your guests would ask, and write down which businesses get named. Run the ten second FAQ check on your own site. And expand your FAQ with the specific, quotable answers travelers are already asking their AI assistants for: prices, policies, and timing.
If your visibility test comes back and a competitor’s name shows up instead of yours, that’s a starting point, not a verdict. Book a strategy call and we’ll help you work out exactly where to start.
