When a traveller asks AI where to book, how often does it say TUI?

What the engines literally say about TUI

These are real, unedited excerpts from this run. The most damaging are not opinions, they are answers where a traveller asked exactly the question TUI should own, and the engine recommended rivals while never mentioning TUI.

For balance: when the engines do name TUI, they do it with authority. The brand equity is real. The job is making the live-search engines retrieve it.

ChatGPT matters most, and it's where TUI is weakest

Ordered by how many real people actually use each engine. ChatGPT leads consumer AI by a wide margin, so it is the number that counts, and it is TUI's softest surface. Claude is shown last only for contrast: it answers from training memory with no live web, which is exactly why it flatters the brand.

Home market strong, biggest market exposed

Splitting the same run by market is revealing. TUI is near-unbeatable in German-language answers, but in the English-language UK market, its largest source of customers, it is barely a coin-flip against the OTAs.

In Spanish, the local OTAs own the answer

A separate, dedicated run of 16 Spanish-language buyer-intent queries, benchmarked against the operators that actually win in Spain: eDreams, Logitravel, Destinia, Halcón Viajes, Viajes El Corte Inglés and the rest. This is a clean, ownable opportunity, and the one I would take first: my home market, my native language.

Spanish-language share of voice

What Spanish travellers actually get told:

And it is a visibility gap, not a product gap

Where TUI wins, and where it disappears

The same brand can be the top recommendation for one question and absent from the next. The cruise line is the clearest blind spot: Marella is nearly invisible when travellers ask AI about cruising.

Who the models name most in this category

Mentions across all engines and queries. TUI leads the named field, but the OTAs (Expedia, Booking.com) and the UK package challengers (Jet2holidays, easyJet holidays, loveholidays) are the ones slipping into answers when TUI is absent.

Who slips in when TUI is left out

For each rival: the number of answers where it was named but TUI was not. This is the recoverable list, ranked. The OTAs top it, which is the structural threat to a tour operator in the AI-answer era.

The pages AI reads before it answers

When the live-retrieval engines cite a source, these are the domains they pull from. Winning GEO means winning the pages on this list, not just tui.com. This is the off-site authority target.

Access is not the problem. Retrievability is.

A quick look at tui.com's public signals. The headline: TUI is not blocking the AI crawlers, so the low live-search numbers are not an access problem. They are a retrievability and authority problem, which is the harder and more valuable thing to fix.

The plan I would put in front of you

Start where TUI is most exposed and where I have a native edge: Spain. I am Spanish, based here, and I speak the language the engines answer in. Win the Spanish market first as a bounded, measurable mandate, then take the same playbook to the UK and DACH. This is the engagement I would run as your external GEO lead.

I have already run this exact method on a real travel brand

Odisea Tours is a Spanish travel company where I ran the full playbook end to end: baseline the share of model, ship the fixes (structured data, llms.txt, an entity/Wikidata footprint, citation content), then re-measure. It is a focused DMC, not TUI's scale, so treat this as proof of method and operator, not of scale. The movement is real, and it is on the engines that matter.

Honest note: the training-corpus channel (Claude, no live web) is the slow 6 to 12 month play and has not moved yet. Every gain above is on the live-search engines, ChatGPT, Gemini and Perplexity, which is exactly where TUI's opportunity is too. Measured 2026-06-16.