AI & Travel · Updated August 2026
Travel is becoming personal in real time.
AI has already changed how guests search, ask and compare — from instant translation at reception to recommendations built around one traveler instead of a mass audience. What it has not changed is who they trust when the money moves.
Originally published in 2024. Substantially updated on 31 August 2026, with new industry data, the current European regulatory position and a revised assessment of what AI in travel delivers today.
AI can adapt communication to language, budget, history and context instead of sending the same message to everyone.
Real-time multilingual communication removes one of tourism’s oldest barriers: two people wanting to understand each other without sharing a language.
Travelers let AI shape the shortlist. Most still refuse to let it book. That gap is the single most important fact in travel technology this year.
AI should remove friction, not people. The translation tool never replaced the host. It simply let his hospitality finally reach me.
— Bourbiza MohamedAI in travel is the biggest shift our industry has seen since the move from paper brochures to online booking — and I learned why the hard way, at a reception desk in eastern Turkey.
The man at the desk was one of the warmest hosts I have ever met, and we could not exchange a single useful sentence. He had only Turkish. I had everything except Turkish. So we smiled, gestured, drew shapes in the air, and lost most of what we were each trying to say.
I have lived that scene more times than I can count: a guesthouse in Iran where the owner sketched me a map by hand, a station in Russia where I simply pointed and hoped, and a family place in Vietnam where the young daughter ended up translating for everyone because school had given her a few English words and the adults had none.
What AI-powered travel personalization actually means
Strip away the buzz and it is simple. AI-powered travel personalization is machine learning reading a traveler’s real context — their language, their history, their budget and the moment they are in — and answering in a way shaped for that one person. Not a brochure. Not a mass email. One reply, fitted to one human being.
My five years running quality at Booking.com taught me one truth that never left me: travelers do not want more choices. They want the right one, handed to them before they have had to ask twice. We tried to do that by hand, with teams and instinct. AI now does it at a scale we could only dream about back then.
Language was always the real wall
Think back to that desk in Turkey. A whole town’s tourism can be quietly throttled by something as ordinary as not sharing a language. The traveler who cannot communicate can hesitate, shorten a stay or simply give up on an enquiry.
When the technology finally caught up, real-time translation was the first thing I built into clients’ guest communication. A guest writes in Arabic. The hotel reads it in Turkish, instantly. The reply goes back in Arabic. No waiting, no fetching someone else to translate.
What follows is direct professional observation rather than measured research, and it should be read that way. On one mid-sized hotel group project, enquiries that used to stall at the language barrier began turning into stays once translation was built into the guest inbox. Nobody hired a translator. I have no independently audited figure for the effect, and I am wary of the percentages that circulate in this field. What I can report is what the front desk reported: they stopped losing conversations.
Where the technology is genuinely working
Real-time multilingual service
Guests and staff communicate across language barriers without waiting for a human translator. The most reliable use in the whole category.
Predictive personalization
Systems use browsing, timing and previous spending to surface offers a traveler is more likely to want. Effective for narrowing, not for deciding.
Dynamic rates
Pricing reacts to demand and local events instead of remaining fixed when conditions change. Mature technology, long before the current AI wave.
Review analysis
AI reads large volumes of reviews and surfaces recurring complaints that teams may otherwise miss. Quietly the highest return of the four.
The 2026 reality check: high adoption, selective trust
Two years after this piece first appeared, the argument has moved. Almost nobody in hospitality now asks whether to use AI. The question is narrower and harder: where to use it without damaging the thing guests are actually paying for.
The most useful evidence I have seen this year comes from Expedia Group’s AI Trust Gap study, conducted by YouGov with more than 5,700 adults in the United States, the United Kingdom and India between 10 and 25 March 2026. Fifty-three per cent said they were comfortable letting AI suggest travel options. Forty-eight per cent said it saved them time and surfaced places they would not otherwise have found. But sixty-six per cent said they would not trust AI to buy or book anything on their behalf, and in the US and UK only eight per cent actually turn to an AI chatbot when planning a trip today.
Read those numbers together and the shape of the market is clear enough. AI has won the research stage and lost the checkout. Travelers are content to let a machine narrow a shortlist. At the moment money moves, they want a name they recognise and someone they can hold responsible if the booking goes wrong.
- Real-time translation in guest messaging
- Review and complaint clustering at scale
- Draft replies a human then edits
- Shortlisting and comparison for the traveler
- Autonomous booking agents
- Granular local detail: opening hours, closures, prices
- “Hyper-personalisation” built on thin data
- Chatbots deployed to cut headcount rather than delay
- Recovery when something has gone wrong
- Judgement calls on refunds and goodwill
- Reading a guest who is upset but polite
- Everything the guest will later describe as hospitality
The pattern on the ground matches the survey almost exactly. Automated answers are excellent until something goes wrong — and then the guest wants a person, fast. A property that has used AI to remove waiting handles that moment well. A property that has used AI to remove staff does not.
What the European rules now require
Since 2 August 2026, the transparency duties in Article 50 of the EU AI Act have applied. In plain terms: a person must be told when they are dealing with an AI system rather than a human, and synthetic audio, image, video and text must be labelled as such. For a hotel, an OTA or a tourism board operating in the European market, that is no longer a matter of taste. Your chatbot has to say what it is.
The heavier obligations moved. Regulation (EU) 2026/1744, the Digital Omnibus on AI, deferred the high-risk requirements for stand-alone Annex III systems from August 2026 to 2 December 2027, and for AI embedded in regulated products under Annex I to 2 August 2028. The labelling duty under Article 50(2) applies to systems already on the market from 2 December 2026.
Most guest-facing travel tools were never going to fall into the high-risk tier anyway, so the deferral changes less for hospitality than the headlines suggested. The disclosure rule changes more. And it points in the same direction as the Expedia data: the constraint on AI in travel is not capability, it is trust. Telling a guest they are talking to a machine costs nothing when the machine is genuinely useful, and costs everything when it is being used to hide the absence of staff.
A warning, from someone who loves this work
Do not automate the warmth out of travel.
Travel is human. It always was. The risk with AI is that we automate so hard we scrub out the warmth — the very thing that made that Turkish host unforgettable even though we could not talk.
Hand the machine the repetitive work, the delay and the language gap. Let your staff be more human, not less.
Where this is going
After 107 countries and a career that began in 1998, my prediction is narrower than it was two years ago, and I think better for it. The language barrier in tourism is going to feel like a relic. Travelers will expect to speak their own language and be understood wherever they go, and properties that cannot meet that expectation will lose enquiries without ever knowing they had them.
What I no longer expect is the fully automated journey that was being promised in 2024. The evidence points elsewhere. Travelers are drawing a line, and they are drawing it in a sensible place: let the machine research, let the human be accountable.
So the real decision for anyone running a property or a destination is no longer whether to adopt AI. It is where to put it — and, just as importantly, where to keep it out. I waited two decades for someone to fix the problem I kept hitting at reception desks. In the end I had to help fix it myself. My advice is still short: do not wait as long as I did, and do not automate the part of the job your guests came for.
FAQ
What is AI-powered travel personalization?
The use of AI to adapt travel communication and recommendations to an individual traveler’s context — language, history, budget and timing — instead of sending the same message to every guest.
Do travelers actually trust AI to book their trips?
Mostly not. In Expedia Group’s March 2026 AI Trust Gap survey, 53% were comfortable letting AI suggest options but 66% said they would not trust it to buy or book on their behalf. AI has been adopted for research far faster than for transactions.
How does AI translation help hotels?
It lets a guest and hotel staff communicate in different languages in real time, removing the friction that stops enquiries, shortens stays or complicates a problem during one.
Does the EU AI Act apply to a hotel chatbot?
The transparency duties in Article 50 have applied since 2 August 2026: guests must be told when they are interacting with an AI system, and AI-generated content must be labelled. The heavier high-risk obligations were deferred to December 2027 and August 2028 and will not apply to most guest-facing travel tools.
Should AI replace hotel staff?
No. Use it to remove repetitive friction, delay and language barriers. The moments guests remember — and the moments when something has gone wrong — still need a person.
Sources
- Expedia Group, The AI Trust Gap — research commissioned from YouGov, 5,700+ adults in the US, UK and India, 10–25 March 2026. Expedia Group partner blog.
- European Commission, regulatory framework for artificial intelligence — Article 50 transparency obligations applicable from 2 August 2026. European Commission, AI Act overview.
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), in force 27 July 2026 — deferral of high-risk obligations to 2 December 2027 (Annex III) and 2 August 2028 (Annex I); Article 50(2) labelling for systems already on the market from 2 December 2026.
