Key takeaways
- Manuela Machner believes the hotel website will disappear in the long run, but will be needed for years to come because the data has to live somewhere.
- AI systems trust sources that provide proof: Booking.com, the Google Business Profile, theme platforms, destination websites, YouTube and even Reddit.
- In Manuela Machner’s analysis for Austria, Gemini cites almost only the Google Business Profile, while Google AI Overviews hardly use it.
- Contradictory information across platforms hurts visibility, because the AI no longer knows what is true.
- According to Google’s AI search, Thomas Hefke was the best speaker at OMR 2026, without ever giving a talk there.
Olga Heuser opens the episode with her own trip to Silicon Valley. AI picked places, restaurants and hotels for her and drove her through San Francisco. Exception or the new reality? That is what episode 2 is about.
Is the hotel website dying?
In the long run, yes, says Manuela Machner, AI expert for tourism. She compares the website to the phone book that every business once had to be listed in. For the next few years, though, a hotel still needs its own website, because the data has to live somewhere. What it will look like and whether it will still be called a website is open.
Thomas Hefke, formerly VP Digital and Head of E-Commerce at Althoff Hotels, explains a hotel to an AI the way you would explain it to your best friend. First the hard facts, then exactly the information she needs to make a good decision.
Why does AI recommend Booking.com so often?
AI systems look for the best answer for the person asking, and for that they need trust. Booking.com has a lot of current data and reviews from real guests, so the systems trust it in particular. For Manuela Machner, though, it is about more than one platform. A hotel has to prove on many levels that it keeps its promises.
Olga Heuser cites an international study from the preparation. More than half of the sources AI cites for hotel recommendations are OTAs. Only 11.8% are websites of independent hotels. According to this study, a strong presence in third-party sources such as YouTube, travel blogs, press articles and Reddit increases the likelihood of a recommendation by a factor of 6.5.
Thomas Hefke also sees an opportunity for specialized hotels. Anyone looking for a romantic hotel on Lake Tegernsee ends up with three to five candidates. For these theme questions, the AI is more likely to trust the individual hotel to say whether it is really romantic than a booking platform.
Which sources does AI use for hotel recommendations?
Which sources count depends on the tool. In Manuela Machner’s analysis of citation sources for Austria, Gemini uses almost only the Google Business Profile, while Google AI Overviews hardly use it. ChatGPT searches via the Bing index and does not read Google Business listings. That is why it pays to connect your profile to Bing once.
- Google Business Profile: Thomas Hefke considers it completely underrated. He recommends keeping offers up to date and uploading your own photos, otherwise Google shows the guests’ pictures.
- Theme platforms: For questions about wellness, families or romance, AI cites wellness and family portals. A bike hotel should therefore be well listed on bike hotel platforms.
- Destination websites: According to Manuela Machner, the websites of tourism regions are strong sources. They often still say the rooms were renovated in 1996. Checking usually costs nothing.
- YouTube, TikTok and Reddit: In her analysis, TikTok ranks seventh among ChatGPT’s sources, and Reddit counts as a trust factor even with just a few posts.
With videos, it is mainly the text that counts. The most beautiful hotel video with the description “Have a great holiday with us” tells the AI nothing.
Why does the data have to match everywhere?
AI compares what it finds on different platforms. If one page says check-out is at 11 am and the next says 12 noon, it doesn’t know what is true. The guests agreed that as soon as there is a contradiction, the AI penalizes it. Thomas Hefke gives the example of opening hours that differ on the website, on Google and on Booking.com.
Reviews matter too. Björn Twardy, Head of IT at Brera Serviced Apartments, stresses how much AI reads from guest reviews. The panel adds another tip. AI also reads the hotel’s replies. A short “Thanks for the great review” wastes the chance to highlight what matters to you.
Is a dedicated app in ChatGPT worth it?
No, says Manuela Machner. She runs trainings every day and has yet to meet a participant who uses an app in ChatGPT. There is also no signal that ChatGPT recommends a hotel more often just because it has an app. Thomas Hefke also sees bigger levers. Anyone who hasn’t done their homework on website, profiles and reviews should start there.
When will guests book directly in AI?
Opinions differ here. Manuela Machner expects guests to book a hotel with an AI agent such as Google Gemini very soon. She believes many people trust their AI assistant more than Booking.com or the hotel itself.
Thomas Hefke rather expects five to ten years. Discovery will happen in the chat, but booking will still take place on a platform or the hotel website. Shopping on Facebook, Instagram or TikTok didn’t catch on either. For decisions involving taste, such as whether you like the bathroom, people want to choose for themselves.
Björn Twardy would book his summer holiday via AI as soon as it is possible. ChatGPT removed its booking feature again in March 2026, Google is working on it. He expects guests to describe their needs instead of choosing a room category, for example a quiet room for a family with noisy kids. At Brera, with 11 properties, 800 apartments and 48 employees, bookings today still come mainly via Google Ads.
What should hotels do now?
Manuela Machner sums it up in two steps. First, decide which question your hotel wants to be the answer to. Then prove in the digital space that it is true.
- Clean up your data. Fill your own website with real content instead of beautiful pictures and little text. Describe room count, target group and offer in concrete terms.
- Check third-party listings. Keep destination websites, your Google Business Profile and theme platforms up to date, with the same information everywhere.
- Test it yourself. Thomas Hefke recommends entering your guests’ questions into ChatGPT and Gemini and seeing who gets cited.
- Focus on two or three channels. Better to maintain YouTube, LinkedIn or Google Business well than to post something on five channels every three weeks.
It is also important not to block search engines and AI crawlers. And once the guest lands on the website, they need to get answers quickly and be able to book.
The full story
Afterwards, Thomas Hefke explains in an interview how he became the best speaker at OMR 2026 without ever taking the stage. He describes which sources he used, why list formats work so well and where he draws the line between optimization and manipulation.
You won’t find that here on the blog, only in the full episode.
Watch episode 2 of Zukunft gebucht now →
Frequently asked questions
Who are the guests in episode 2 of Zukunft gebucht?
Manuela Machner, AI expert for tourism, Thomas Hefke, formerly VP Digital and Head of E-Commerce at Althoff Hotels, and Björn Twardy, Head of IT at Brera Serviced Apartments. The hosts are Olga Heuser (DialogShift) and Anton von Verschuer (RateBoard). More about the format is in the article on the Zukunft Gebucht industry talk.
Does a hotel still need its own website in the age of AI?
Yes, at least for the next few years. AI needs a place where the hotel’s data lives. The website should have concrete, factual content and match the information on other platforms.
How does a hotel become visible in ChatGPT and Gemini?
With clear data on its own website and proof in third-party sources such as the Google Business Profile, review portals, theme platforms and destination websites. All information should match. Gemini relies mainly on the Google Business Profile, ChatGPT searches via the Bing index.
What was episode 1 about?
Episode 1 was about what makes a hotel distinctive. The summary is in the article What makes a hotel distinctive when the fifth star no longer does.
About the hosts
RateBoard
RateBoard develops revenue management software for hotels. The system analyzes booking data, demand and market prices and automatically calculates the best room rates. This helps hotels grow revenue and spend less time on pricing. RateBoard has been on the market for over ten years and is part of the Zucchetti Group.
DialogShift
DialogShift is a Berlin-based software company for AI-powered guest communication in hospitality. With Chat AI, Phone AI, Email AI and Journey Messaging, hotels answer guest requests around the clock on every channel, increase direct bookings and take workload off their teams.
