AI Visibility for Travel Agencies and Tour Operators

Travel prompts are long, specific and stacked with constraints. That works in your favour if your pages state the specifics plainly, and against you if they don't. Here are realistic examples of what buyers type:

• Plan a 10-day small-group trip through northern Portugal in May, under $3,000 per person, with walking and wine tastings.
• Which tour operators run guided Kilimanjaro climbs in February with a 90%-plus summit success rate?
• Compare G-style budget tours and premium private tours for Japan in cherry blossom season.
• Is there a family-friendly Iceland tour that includes the Golden Circle and leaves from Reykjavik in July?
• What is the cancellation policy for [operator name]'s Patagonia trek?
• Is [travel agency name] a licensed and bonded agency, and is my money protected if they fold?
• What's included in [operator name]'s Morocco desert tour: flights, meals, tips?

Notice the split. The first four are discovery prompts where you either get named or you don't. The last three are verification prompts where a traveler already knows you and wants a fact. Wrong answers to verification prompts cost bookings you had already earned.

The engines also rewrite these prompts. OpenAI says ChatGPT search typically rewrites a query into one or more targeted queries, so the exact wording a traveler types is not what gets searched. Treat any prompt list as a sample, not a map of real retrieval.

Travelers use several assistants across a booking cycle that can run for months: inspiration in ChatGPT, comparison in Google, a final check in Perplexity. Each engine works differently, as of October 2026:

• ChatGPT searches the web at the model's discretion. When it doesn't search, it answers from training data with a cutoff. OpenAI's model docs list a February 16, 2026 cutoff for the GPT-5.6 models and August 31, 2025 for the Instant path ChatGPT uses. An unsearched answer can describe last season's departures.
• Google AI Overviews and AI Mode only use pages that are indexed and eligible for a snippet in Google Search. Google says there are no extra technical requirements. Googlebot renders JavaScript, so widget content is less of a risk here.
• Gemini, Perplexity and Claude each run their own retrieval, governed by their own crawlers and settings.
• Microsoft Copilot grounds on the Bing index, so Bing Webmaster Tools tells you something Google Search Console doesn't.

Measure each engine separately. A blended score hides the fact that one engine may describe you well while another gets your dates wrong. Our guide on how to check if your brand is visible in AI search covers the method, and measuring AI share of voice shows how to compare yourself against other operators on the same routes.

AI platforms get specific facts wrong about specific brands. For travel businesses, the risk sits in facts that change every season or vary by departure:

• Departure dates. A model trained on last year's calendar, or a crawler that saw an empty widget, can state dates that no longer exist.
• Prices and currency. Per person versus per room, single supplements, and whether the price is from-pricing or fixed.
• Inclusions. Flights, internal transfers, meals, park fees, tips and equipment. Travelers ask about these directly, and a wrong answer creates a complaint before the trip starts.
• Cancellation and deposit terms. Refund windows, deposit amounts and change fees are easy to misquote when they live in a PDF or a checkout step.
• Licensing, bonding and trade memberships. Whatever licence, bond or consumer protection scheme applies in your market, state its name and number on a fetchable page. A vague answer here reads as a warning to the traveler.
• Group size, fitness level and age limits. These decide whether a family or older traveler books.

The one positive experimental signal for on-page content is simple: state the specific fact in a sentence on a page that can be fetched. "The 2027 season departs every Saturday from 3 April to 25 September" gives an engine a sentence it can quote. A date picker gives it nothing.

Most tour sites load availability, departure dates and live prices through a booking widget or an embedded third-party engine. That content often appears only after JavaScript runs in the browser.

Vercel's crawler study found that ChatGPT and Claude crawlers fetch JavaScript files (11.50% and 23.84% of requests) but don't execute them, so they can't read client-side rendered content. Independent tests in June 2026 found the same for OAI-SearchBot, Claude-SearchBot, PerplexityBot and several others. Googlebot is the exception: it renders.

Say you run small-group walking tours in Portugal. Your tour page has a strong description, then an iframe calendar from your reservation system. To a non-rendering crawler, the page has no dates, no price and no availability. The engine fills that gap from wherever else it can: an old marketplace listing, a blog post from two seasons ago, or its training data.

The fix doesn't require replacing your booking system:

1. Fetch your tour page with JavaScript disabled and read what's left.
2. Add a plain HTML block above the widget with the season's departure dates, the from-price, inclusions and a link to the full terms.
3. Put cancellation terms and your licence or bond details in HTML on a page you link from every tour, not only in a PDF or checkout step.
4. Repeat the check after every redesign or booking-engine change.

Access is binary. If a search crawler can't fetch a page, nothing on that page reaches the answer.

robots.txt is per bot

OpenAI's crawler page says each setting is independent. GPTBot governs training use. OAI-SearchBot controls whether you appear in ChatGPT search answers. A blanket block on AI bots removes you from ChatGPT search. A pattern that allows search while declining training looks like this:

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

There is no robots.txt token for Google AI Overviews or AI Mode. Googlebot access is the control there.

The firewall can override robots.txt

OpenAI's help centre says eligibility needs both robots.txt permission and a host or CDN that allows traffic from its published search bot IP addresses. Bot protection on booking sites is often aggressive, because scrapers target fares and availability. That protection can return a 403 to a legitimate AI crawler. Cloudflare also changed defaults on 15 September 2026: Training and Agent crawler categories are blocked by default on ad-displaying pages for new domains, new sites on existing accounts and existing free-tier customers, while Search stays allowed. Check your settings rather than assuming.

The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away.

Accidental suppression

noindex, nosnippet, max-snippet:0 and data-nosnippet keep content out of Google's and Bing's AI answers. Templates sometimes apply them to whole tour pages or terms sections by mistake. Search your templates for them.

For many itinerary prompts, engines cite marketplaces, review platforms, destination guides and travel publications instead of operator sites. The evidence shows a strong association between being named on the sites engines cite and appearing in the answer. It doesn't promise an effect, but it does mean those listings are often the version of you a traveler reads.

Check these, in this order:

• Tour marketplaces and booking platforms where you list departures. Old seasons, retired tours and outdated prices linger here longest.
• Review sites. Look at how your business name, tour names and location appear, and whether old reviews describe a tour you've since changed. See whether online reviews affect AI recommendations for what the evidence does and doesn't support.
• Comparison and "best tours" articles that engines cite for your destination. Our page on whether comparison pages help AI recommend you covers this.
• Trade body and licensing registers. If a register lists your licence or bond, make sure the business name matches your site exactly.
• Partner agency pages that resell your tours with their own copy.

To find which sites matter, run your discovery prompts and record every cited URL. Fix the ones that repeat a wrong fact before chasing new listings.

Week 1: baseline

Write eight prompts: four discovery, four verification, using your real routes and tour names. Run them in each engine, logged out, and record what each says about dates, price, inclusions, terms and licensing. The free checker on this site gives a quick snapshot for three buyer questions.

Week 2: access

Check robots.txt for OAI-SearchBot, PerplexityBot, Claude-SearchBot and Googlebot. Test whether your firewall or CDN returns a 403 to those crawlers. Search templates for noindex and snippet controls. Confirm index status separately in Google Search Console and Bing Webmaster Tools.

Week 3: facts in HTML

Fetch each top tour page with JavaScript off. Add an HTML block with season dates, from-price and inclusions. Publish one terms page covering deposits, cancellation and licensing or bonding, in plain text.

Week 4: third parties and re-test

Update or retire outdated marketplace listings and partner pages. Respond to review profiles where your tour details are stale. Re-run the week 1 prompts and note which wrong facts persist and where engines still source them.

If you're deciding who should own this, read whether you need an expert or can do it in-house and whether your SEO agency should already handle it.

Do I need to replace my booking engine to fix AI visibility? No. The problem is that dates and prices appear only after JavaScript runs, which most non-Google AI crawlers can't see. Keep your booking engine and add a plain HTML summary above it: departure dates for the season, the from-price, inclusions and a link to terms. Then fetch the page with JavaScript disabled to confirm the facts are there.

Should I allow GPTBot so ChatGPT recommends my tours? GPTBot governs whether your content can be used for training. It doesn't control ChatGPT search. OpenAI documents OAI-SearchBot as the crawler for ChatGPT search answers, and says each setting is independent. You can allow OAI-SearchBot and disallow GPTBot. Allowing a crawler buys eligibility only, not a recommendation.

Why does ChatGPT still quote last year's departure dates? Two common causes. When ChatGPT answers without searching, it relies on training data with a fixed cutoff, which can predate your new season. When it does search, it may cite an old marketplace listing or article because your own page hides dates in a widget. Publish current dates in HTML and update the third-party pages it cites.

Can I make AI assistants recommend my tour company? No one can promise that, and claims of guaranteed AI rankings have no evidence behind them. What you control is access and accuracy: crawlers can reach your pages, current facts are in fetchable HTML, and listings on other sites are correct. Which operators a model names for a given prompt is decided by the engine.

How often should a tour operator re-check AI answers? At least at every season change, after any booking-engine or site redesign, and whenever you change prices, inclusions or cancellation terms. Engines change models and retrieval behaviour too, so a monthly check of your verification prompts is a sensible rhythm for businesses with frequent departures.