More and more guests start their research through AI systems such as ChatGPT, Gemini and Claude. It is no surprise, then, that a growing number of tools and providers promise to make your hotel more visible within these models. At the same time, this is a young field in which much is still unclear about what truly has an effect. At Becurious we follow these developments closely and advise our clients on what is worthwhile and what is not.
AI models draw their information from the broader structure of the web. Well-implemented structured data (schema.org) helps these models interpret facts about your hotel correctly: location, facilities, room types, restaurants and opening hours. That is why we include structured data as standard in every website configuration.
It is important to know that more is not automatically better here. When an external party also adds structured data through a separate script, duplication occurs. That can actually confuse AI models and have the opposite of the intended effect. One consistent, well-maintained source works better than several sources that overlap.
A common piece of advice is to place an llms.txt file on your website, comparable to robots.txt but aimed at AI models. The idea sounds logical, but the reality is different: the major AI providers do not use it at this time. Research across hundreds of thousands of domains shows no measurable effect on how often a website is cited by AI.
That is why we deliberately do not place this file yet. It takes development time without demonstrable results, and an outdated or poorly maintained file can even work against you. As soon as the major models do start supporting llms.txt, we will of course make it available right away. We then place and manage it as a standard part of our own website technology, directly from the CMS, so that it stays in sync with the rest of the site and is always up to date.
The biggest gains in AI visibility do not come from a single technical file, but from the fundamentals:
It is precisely on those external sources that we often find outdated information in practice, about ownership, team, awards or offerings. That is exactly what AI models pick up and pass on.
The AI landscape changes quickly. What has no effect today may be standard within six months. We continue to track these developments, test what works and advise our clients on the right choices. Have you received a request or recommendation from an external party and are you unsure whether it is worthwhile? Feel free to run it by us.