AI guest communication on the website: How Regensburg Tourismus GmbH answers its guest inquiries as a DMO

From skepticism to a three-year partnership

RTG has been a branchly customer for over three years and was thus one of the very first destinations to launch a generative AI chat on the German market. Before that, the team was explicitly skeptical. “I don't want yet another person promising me the moon,” Molski describes her attitude at the beginning of the collaboration. In the end, a simple, consistent idea convinced her: the knowledge base is the website itself. Everything that the Chat or Search outputs comes from RTG's curated pages and is therefore researched and correct. Maintain once, both interfaces learn along with it.

The implementation proceeded step by step and grew with the requirements. First the Chat went live, then the search bar was added to the homepage, later the telephony connection, then the Regiondo integration for city tours and tickets, and currently the digital guest folder is being created. At every step, the basic principle remained the same: no second system, no second knowledge base, but one data foundation that feeds all channels. Beyond its own website, the knowledge base was also expanded to include secondary and external sources that were reviewed by the RTG team, for example on parking options or the surrounding area.

The collaboration itself has become an ongoing dialogue over the years. RTG provides requirements from guest contact, branchly implements them, and both sides have been learning from each other for three years. This is how Easy language in the Chat came about, for which branchly and RTG worked with a specialist to develop the implementation, a project with which the partnership won the ADAC Tourism Award Bavaria in 2025. As a destination without the budgets of Cologne, Hamburg, or Berlin, RTG approaches digitalization in deliberately small but visible steps, and the hurdles of municipal IT, from firewalls to security precautions, are honestly part of everyday life.

One system for Chat, Search, and telephony

At the center are Chat and hybrid Search, which guests use equally. On the homepage of tourismus.regensburg.de, there are therefore two entry points: the familiar search bar for guests who expect “no robot,” and the chat bubble for everyone who prefers to ask in dialogue. Both are used about half the time. “Luckily, we implemented both, otherwise we probably wouldn't have so many successful inquiries,” says Molski. “All information that the Chat or Search outputs on the website is researched and correct.”

The Chat automatically detects the language of the request and answers in the local language, from English to Dutch and Italian to Czech. Easy language is also integrated. The destination's booking systems are directly integrated into the dialogue: via the Regiondo integration, guests receive city tours with date, price, and direct booking link as selectable cards in the Chat, plus current events and weather information. In total, 33 curated sources feed the knowledge base, from UNESCO World Heritage to parking options to walks through the old town.

In addition, a telephony connection with six languages runs outside opening hours. Around 200 calls per month are successfully handled by the AI; requests that remain open end up as a summary with the team, which then calls back in a targeted way the next morning. The decisive point is that the calls flow through the same data foundation. “We pull all phone scripts through the platform and thereby see different user behavior between guests who search on the website and guests who call,” says Molski. “Are they the same topics? Are they more complex requests?” RTG learns from every channel without maintaining a second system: “At the end of the day, as a DMO, I am satisfied because I had no extra effort.” As a next step, a digital guest folder is currently being created that picks up guests from hotel check-in with tailor-made programs from the same data.

Figures from guest contact

The figures come from RTG's ongoing application:

Metric	Value	Time period
Fully answered	94.6%	Last 12 months
Interaction rate	7.2% of website visitors interact with the AI interfaces	Last 30 days
Clicks on ticket and booking pages	5,022 on event tickets, 2,765 on the booking portal for overnight stays	Last 12 months
Languages in Chat	10+ detected (93% German, 7% international)	Last 30 days

At least as revealing as the metrics themselves for RTG is what shows behind them. Guests no longer search only for opening hours: they ask about events on the weekend, about parking, about the perfect day trip with children, and the AI leads them directly to ticket purchase. Almost 8,000 clicks on event tickets and the booking portal for overnight stays in twelve months show that the dialogue regularly turns into bookable actions. The Tourist Information team is relieved of standard inquiries and has time for the cases where personal interaction matters. And the inquiries themselves have become data material from which RTG builds its marketing: “Regensburg is beautiful and a UNESCO World Heritage site. But perhaps that is not what we need to put at the center, but rather that it is a vibrant city,” says Molski.

What other destinations can learn from this

Perhaps the most important takeaway lies beyond the individual metric. “We are all so rich in data and do far too little with it,” Molski describes an industry that collects inquiries, phone calls, and visitor focus areas without evaluating them. In Regensburg, insights from Chat, Search, and telephony now flow directly into the print magazine, website content, and the destination's own discoverability in AI portals.

Two topics will shape the coming years. First, visibility in AI itself: guests increasingly plan trips in dialogue with language models, and the destination that cannot be found there loses the first contact. Second, the destination's own website remains the reliable reference point. It is no longer the place of decision, but the place of confirmation, and, according to Molski, it must be able to “speak with AI.” Together, both create the future vision: a website that, as a curated source, answers both guests and AI systems, and an on-site team that delivers the encounter that no machine can replace. The insider tips, “which corner you have to turn in Regensburg,” remain human knowledge.