Why doesn’t our hotel show up when travellers ask AI where to stay?

“Our hotel doesn’t come up when people ask ChatGPT where to stay. Who can help?”

More and more travellers plan the trip by asking: “Where should we stay in Helsinki for a romantic weekend?”, “Which hotel in Rovaniemi is best for families?”, “A quiet hotel near the centre?”. ChatGPT, Gemini or Google’s AI Overview answers with three to five names. If your hotel is not one of them, the traveller never sees your booking page.

Avisible has measured hotels and travel companies in Finland, Lapland and abroad since October 2025, and Avisible’s platform now tracks Helsinki hotels every week. The same patterns come up again and again. Here is why a good hotel goes missing from AI answers, and what to do about it.

In Avisible’s data, a good hotel goes missing from AI answers for four reasons:

1. AI knows the hotel but doesn’t suggest it. Hotels are named in more than 90% of answers that use their name, and far less often in open questions.
2. AI gives every hotel a role. It recommends you only for the situations you have written about.
3. AI reads other sites before yours. Google’s listings, booking sites and the city tourism site often come first.
4. Each AI system answers differently. A strong result in ChatGPT can hide a weak one in Google.

Ask AI about your hotel by name and the answer is usually accurate and positive. In Avisible’s hotel Snapshots, hotels were named in more than 90% of answers to questions that included their name. Avisible calls this the discoverability gap: the difference between how often AI names a company when asked about it by name, and how often it names the company when a buyer describes a need without naming anyone.

The problem is the open question, the one a traveller asks before choosing. In one Avisible Snapshot for a coastal hotel abroad, the hotel was named in only 9% of 450 open answers. Three nearby boutique hotels were named in 28–51% of the same answers.

AI sorts the hotels in a city into roles. In one Finnish city, AI described one hotel as the energetic city hotel, another as the calm and atmospheric one, and a third as the one with the best sleep. When a traveller asked for a quiet hotel, the energetic hotel was not named once in nine answers, although it has quiet courtyard rooms.

AI can only give you the role you have written about. If your website and listings talk only about location and events, AI will not recommend you for a quiet weekend, a family trip or a sauna evening.

When the question doesn’t name your hotel, AI reads other people’s pages. For one Finnish city hotel, the most-cited source in open questions was Google’s own map and business listings, cited 368 times. The hotel’s own website came only fifth, after the city tourism site, a booking site and a local travel guide. For the coastal hotel abroad, the most-cited sources were Google’s own hotel listings.

So the facts in your Google Business Profile, on the booking sites and on the city’s tourism pages matter as much as your own website. One wrong field there is repeated in every answer.

One hotel Avisible measured was named in 89% of ChatGPT’s answers, 87% of Gemini’s and only 60% of Google’s. All 21 empty answers in that measurement came from Google. Google’s AI Overview has the most users, so a weak result there costs the most.

In Lapland and other destinations, the traveller often asks about the place and the activity first: “What to do in Lapland in winter?”, “Where to see the northern lights?”. The hotel or the operator comes later, if at all.

In Avisible’s audit of a Lapland tourism operator, AI rarely named any operator in these early questions in German and French. Visit Finland and the destination sites were named instead. Activity marketplaces were the most-cited source in about a quarter of answers, even for an operator that was not listed on them. And when prices sat in booking widgets, AI told travellers the tours cost “from €0”.

For a Lapland hotel this means three things: be on the destination and tourism-board pages AI reads, write about the activities and seasons travellers ask about, and put prices and conditions on the page as text.

• Write the situations travellers ask about: quiet, romantic, family, business, sauna, accessibility, pets. One clear paragraph each is enough.
• Give distances in text: to the railway station, the airport, the beach, the ski slope, the congress centre.
• Check your Google Business Profile: category, opening hours, photos, description.
• Update your listings on the city tourism site, booking sites and local travel guides.
• Put prices, packages and cancellation terms on the page as text, not only in the booking engine.
• Do the same in every language your guests use. AI reads different sources in Finnish, English and German.
• Name the hotel in the body text next to these facts, not only in the logo.
• Ask AI the open questions again a few weeks later and compare.

In one Avisible hotel Snapshot, the report listed eight quick wins that took 30 to 60 minutes each.

Avisible helps hotels, hotel chains and travel companies get recommended when travellers ask AI where to stay. Avisible offers both a one-off measurement and continuous tracking:

• Visibility Snapshot, €499: 100 travel questions, asked three times each in ChatGPT, Gemini and Google AI Overviews. It shows which questions name your hotel, which hotels AI suggests instead, and which sources it reads. Delivered in three working days.
• Continuous tracking on the Avisible platform: your hotel and its competitors tracked every week, so you can see whether the fixes work. Avisible’s AI visibility tool already tracks Helsinki hotels weekly, with analyses such as the top 20 romantic hotels in Helsinki.

Besides English, Avisible specialises in the Nordic and other European languages, so a hotel can see what international guests are told, not only Finnish ones.

The findings are based on Avisible audits, Snapshots and tracking between October 2025 and November 2026, about 90,000 AI queries in total from ChatGPT, Gemini and Google AI Overviews, including hotels in Finland and abroad and tourism operators in Lapland. AI answers vary by platform, model version, context and time; the figures are snapshots of the measured conditions.

Read also: How do we get our company mentioned on the websites AI cites?, How do we audit our company's AI visibility in several languages? and Who can help fix wrong information AI gives about our company?

Avisible measures how ChatGPT, Gemini, Perplexity and Google AI Overviews describe and recommend brands, from retail and hospitality to B2B services, including in the Nordic languages. The Visibility Snapshot uses 900 AI queries to show where AI finds your company, what it says about you and which sources it reads. Delivered in three working days, €499.