Hospitality
Conversational selling in hospitality: from the in-room QR to the event kiosk
Case study: an in-room QR, a virtual host profiling guests in 5 languages, a call in under 30 seconds, and an event kiosk with per-promoter attribution.
<30 s
from form submission to the voice agent's call, in the guest's language
System capacity figure, not a client business result.
35
videos produced: 7 pieces in each of 5 languages
Pipeline capacity figure, not a client result.
7 days
of continuous kiosk operation at a golf tournament
Verifiable on the ground; we don't publish client business metrics.
An international hotel chain with resorts in the Caribbean asked us for something more concrete than “add AI to the hotel”: turn the guest’s moments of attention into sales conversations. We built an end-to-end experience that starts at a QR code — in the room or on an event floor — and ends where sales actually close: in a voice call. In between, a virtual host on video who welcomes the guest in their own language, a quiz that builds their traveler profile, a different pitch per profile, and a voice agent that dials in under 30 seconds after the form. And at a golf tournament, the other half of the story: a kiosk with a conversational avatar that ran for seven straight days, with a unique QR per promoter and an event log that attributes every booked meeting to the person who sparked it.
The starting point: plenty of attention, zero conversation
A hotel has what almost no other business gets: the customer physically inside, with time on their hands. What that customer used to find was an acrylic stand with a QR that opened a PDF. A PDF asks nothing. It can’t tell the family with two kids from the anniversary couple, and it’s written in English — little help to someone who flew in from Kyiv or Tbilisi.
Events had the equivalent problem: promoters handing out material, one generic QR on a banner, and contact data typed up by hand days later. When a meeting got booked two weeks on, the answer to “where did this customer come from?” was “the tournament.” Nothing more.
Inside the hotel: a host who speaks your language
The QR opens a language picker — Spanish, English, Chinese, Ukrainian, Georgian — and there she is: a virtual host with a local accent and subtitles, welcoming you on video. Not a menu of buttons; someone talking to you.
Then she asks what kind of trip this is. Here’s the detail guests never consciously notice but always feel: the answer options surface over the video at the exact instant she speaks the question, synced word by word. The answers build a profile, and she follows with a micro-pitch made for it — a family hears a different pitch than a couple.
The close isn’t “thanks, we’ll be in touch.” It’s a short form and, in under 30 seconds, the guest’s phone ringing. On the line, a voice agent that already knows their language, their profile, and what caught their interest. Pick Georgian on the first screen and the call arrives in Georgian. The thread never breaks.
At the event: a kiosk that converses and a QR that attributes
Across the tournament’s seven days, a touchscreen kiosk with an avatar greeted players in real-time conversation. The roster said “G. Sanchez Rdz.” and the player gave another name; fuzzy name matching found him anyway, and check-in flowed with no staffer hovering with a tablet. The live leaderboard ran on the same screen, and that hook did the heavy lifting: players kept coming back to check standings, and every return was another conversation.
Each promoter on the team carried their own QR code. Scanning it didn’t open “the website” — it opened a journey signed, from the first second, by the person who sparked it. Everything that followed — the visit, the form, the call, the WhatsApp thread, the meeting — was written as a chain of events tied to that scan, impossible to retouch afterwards. On that same foundation, A/B/C tests of the digital experience ran with weighted traffic assignment per variant.
The factory behind it
Seven video pieces per language — one hook, five micro-pitches, one closer — across five languages: 35 videos. With traditional production, that’s weeks of casting, studio time, and voice talent per language. Our AI avatar video factory produced and deployed them in hours, and it’s the same factory that produced the hero video on this site’s homepage. Changing a script, adjusting a pitch, or adding a language iterates like software, not like a TV commercial.
The numbers we do publish
We don’t publish our clients’ business metrics. What is ours and measurable: a call in under 30 seconds after the form, 35 videos in 5 languages produced and deployed in hours, and seven days of continuous kiosk operation under real event conditions (system capacity figures, not client results). The qualitative part carries equal weight: the guest never repeats information, never gets switched to English halfway through, and every meeting from the event was attributed to a promoter beyond dispute.
Where is your moment of attention?
If your business has a spot where the customer is already present and nobody converts them — a room, a sales floor, an event — book a 30-minute session. We’ll walk you through the full experience, QR to phone call, and you’ll leave with a diagnosis of whether it fits your case.