What a luxury travel recommendation actually depends on

More than the preference profile. A luxury travel recommendation depends on the client's taste, the live weather where they are going, and real-time safety conditions in the specific neighbourhood, held together at the moment of the suggestion. Sofiane Ghorbel, founder and CEO of AL Hospitality Group and more than twenty-five years into luxury hospitality, named three inputs no stored profile supplies.

What does a preference profile miss?

Everything that changed after the profile was written. Taste is the part a profile does hold, and Ghorbel starts there:

"They will provide us with some example of suggestions of restaurants, for example, museums, because he used to love arts — only arts, arts, arts, but not history."

The other two inputs are true only today, which is why a stored record cannot carry them.

Why is the weather not a minor detail?

Because in luxury travel the weather does not degrade a recommendation, it invalidates it. Ghorbel's example is a Paris terrace:

"Terraces in Paris — you cannot go to a terrace in Saint-Germain-des-Prés, Les Deux Magots for example, when it's raining. You cannot have fun on the terrace, because the fun is the terrace, it's not inside the cafe ... people want to be seen. And to be seen is on the terrace."

A tool that treats weather as a comfort variable will keep producing answers that are technically correct and professionally useless. The same restaurant is the right answer on Tuesday and the wrong one on Wednesday.

What does real-time safety mean in this workflow?

Conditions that appear with no warning in one part of one city. Ghorbel:

"Depends if they have riots in this particular area ... riots sometimes happen suddenly. We don't expect [it] — after a soccer game for example, there are riots like last month, it happened in Paris. So they don't need to be in a specific area where the riots are."

This is a live feed requirement rather than a country risk rating. The granularity that matters is the district, and the freshness that matters is hours.

What should the tool do with all of it?

Combine it. Ghorbel's own summary of where AI earns its place:

"This is the strength of the AI: to get all those information and all those data from everywhere. Makes it like a cocktail and give us a great cocktail."

Galina Fendikevich extended it to the technical requirement: a multimodal, real-time view drawing on social platforms and live reviews rather than static text, narrowed to what this industry needs — political and social conditions in a city, restaurants, weather, and vetted experience providers, combined with the advisor's own knowledge of the client.

Where should the AI stop?

At the judgment. Ghorbel draws the boundary himself, and it is the sharpest line in the interview:

"We know AI lacks taste. Because luxury is not information, luxury is judgment, luxury is knowing."

"Which hotel fits a family, which destination fits an occasion, which experience fits personality, when not to recommend something. So the most valuable hospitality skills remain human judgment."

When not to recommend something is the item most product teams would never build for. Every recommendation system ever shipped optimises for producing an answer, and Ghorbel names restraint as a professional skill.

What about the discovery stage?

He wants immersive discovery there, and says plainly that it does not exist:

"In this step of discovery [and] curation we can use ... augmented reality, the AR and the VR. We don't have this tool now. But in the future I hope that we will have it — when the goggles, the Apple goggles, will be more democratized and cheaper."

What he wants clients to do with it is meet people, not view rooms: "They can discover the destination, or two or three destinations, with the goggles. They can even meet people there in the destination" — the team, providers, a tour guide, or the hotel concierge.

What to check before buying a recommendation tool

Artificial Reality's read of Ghorbel's requirements, for an operator evaluating vendors:

• Does it read weather at neighbourhood level, on the day, or only climate averages?
• Does it carry live local conditions, including unrest that started last night?
• Can it decline to answer, or does it always produce a suggestion?
• Does it take your own knowledge of the client as an input rather than replacing it?
• Is it built for hospitality, or a generic assistant pointed at travel?

Ask a vendor to run one live itinerary in a city where the weather turned this week. The demo will answer all five questions at once.

Ghorbel's own boundary is the standard to hold them to: "AI does not replace expertise, it accelerates expertise."

Where this comes from

S1E3: Why AI keeps failing in luxury hospitality — the full interview with Sofiane Ghorbel. Listen or watch: YouTube, Spotify or Apple Podcasts.