Results
Real cases, not promises
AI projects in production today, told honestly: what problem they solved, what we built and what changed. No client names, with numbers.
What already works, with numbers
In production today, facing real clients. Anonymized, no names, no hype.
A virtual concierge that never closes
7 languages
24/7 · 35-40 chats a day · ~30% in another language
A conversational assistant embedded in the hotel's website that answers guests in their language — up to 7 — day and night, grounded in the hotel's real information (RAG). It runs 100% on the hotel's server: not a single conversation leaves home.
From hearing it from the client to hearing it first
3 days → <15 min
detection time on a real incident
Sentinels that check every few minutes that critical services are up and alert instantly via Telegram. On a real incident, the detection window dropped from up to 3 days to under 15 minutes.
The competition's price, every morning with your coffee
OTA parity
AI-written reports, daily
It gathers competitors' prices every day and generates automatic reports: parity between the direct price and the OTAs', positioning and sharp price moves. Hours of manual tracking turned into one clear report.
Get ChatGPT to recommend your business
Referred by AI
hotels and clinics, with measurable audits
Our own methodology and tool (GEO) so AI engines know and recommend your business. Already with real clients: hotels, and psychology clinics that get patients referred by ChatGPT. With before-and-after audits, not promises.
Before the guest notices
24/7 monitoring
dozens of rules · mobile alerts via Telegram
Centralized observability of servers, network, sites and backups, with dozens of alert rules (disk full, failed backup, SSL about to expire…) reaching your phone sorted by severity. Proactive detection before the client notices.
Bespoke apps, under your control
Overbooking caught on its own
in-house portals, not exposed to the internet
A sun-lounger portal with live occupancy and automatic overbooking alerts, and parking management tied to the booking. Inside the hotel's infrastructure, integrated with the rest of the systems and with an audit log.
In development, and we say so
Projects in research or development — not yet in production. We tell them as what they are.
Direct-sale chatbot
An assistant that converts the visitor who hasn't booked yet: live availability and price, handles objections and drives the direct booking, winning ground back from the OTAs. In development.
Computer vision against food waste
AI cameras to measure buffet and kitchen waste and adjust purchasing toward zero waste. Hardware available; in development.
Intelligence over the PMS
An AI layer that reads (never writes) the hotel management system for occupancy forecasting, pricing support and no-show prediction. Connection proven; models in development.
Fully-local voice assistant
Talk to the AI and have it answer by voice, without the audio leaving your server. Working prototype.
Want these results in your operation?
We'll tell you what we'd automate first in your case and which metric would tell you it works.