The biggest AI risk for a hotel? Not using it

Standing still is also a decision

Your guests already use AI to plan their trips and choose where to stay. Your competitors are starting to use it to answer faster, prepare quotes and follow up on every request. A hotel that waits "until AI matures" isn't avoiding a risk; it's choosing a different one: falling behind while others learn. And learning takes time, so starting later rarely means catching up faster.

The second risk: using it with nobody in charge

Using AI badly rarely means choosing the wrong tool. It means letting it answer guests, publish content or make decisions with no one checking. AI makes mistakes, and it states a wrong answer with the same confidence as a right one. The model Tomeu described on Hosteltur TV is simple: agents that search, agents that verify, and a person in charge. AI prepares and checks the work; a person decides, and is responsible for the result.

AI isn't here to cut jobs

This is the first fear in most teams, and in a hotel it's misplaced. The time AI frees up doesn't disappear: it goes back to the guest. Less time copying data between systems and answering the same email for the tenth time means more time to look after guests, to upsell, and to follow up on a group quote before a competitor does. The hotels that get the most out of AI will use it to serve and sell more, not to cut staff.

How to start without falling into either risk

• Pick one repetitive, measurable task, such as answering the most common guest questions or preparing a weekly report. We explain how to choose it in Where should a hotel actually start with AI?
• Name the person in charge of the result before you start.
• Measure before and after, so you know whether it works and can defend the next step.
• Build it into daily work from day one, not as a side experiment. That's why most AI projects never make it past the pilot.

The risk is not using AI. And then, using it badly.