Reads live PMS data, checks availability against policy, prices offers in real time, and writes the action back to your folio. Across 200+ languages. Hands off to staff with full context when judgment matters.
.png)
Built for real guest messages — half-formed, typo-ridden, switching between English and the guest's native language mid-sentence. The AI classifies intent, not keywords. Four ways to ask for a late checkout resolve to the same request, in the same second.
Before the AI says a word, it reads the reservation, profile, folio, availability, and your knowledge base. Every policy, rate, and room assignment comes from your systems — not from a guess.
The difference between a chatbot and an operator is the verb. Viqal books services, charges the folio, creates housekeeping tasks, and updates guest profiles — all with confirmations sent back to your PMS, not emailed to a queue.
A chatbot answers questions. An AI operator for hotels does the work — reads live PMS data, executes folio charges and bookings, writes confirmations back, and hands off to staff when judgment matters. That's the difference between a front-desk shortcut and a second shift on the team.
Judgment calls go to humans. When confidence is low, the task is out of scope, or a guest is upset, the AI escalates to Team Inbox with the classification, reasoning, and a suggested reply staff can send or edit in one click.
Scripted chatbots match keywords to programmed replies — fine until a guest deviates. General-purpose LLMs construct answers based on context. An AI operator for hotels combines: intent classification over keywords, plus live PMS and knowledge-base reads before every reply and it hands off when a guest needs a human.
Anything you've opted into: booking services (spa, transfers, dining), extending checkouts, charging the folio, updating guest profiles, creating housekeeping or maintenance tasks. Each action is logged with a PMS confirmation. You set guardrails per action — amount caps, approval thresholds, offers only allowed in certain windows.
Every reply is grounded in structured PMS data and your knowledge base. The AI can't quote a rate your PMS doesn't publish or promise a policy that isn't documented. If a guest asks about something outside those sources, it says so and hands off — transparent, auditable, no confabulation.
A structured editor in the dashboard — policies, hours, amenities, FAQs, brand-voice rules, escalation triggers. Edits go live immediately and version themselves. Most properties seed it in a half-day kickoff with Viqal, then edit in place as their ops evolve. No retraining, no waiting.
On low confidence, explicit guest request for a human, urgent maintenance, medical language, complaints, VIP flags, or actions outside its permissions. Staff get the full reasoning trace, intent classification, and a suggested reply — so picking up a handoff is a one-click send, not a re-read.
.avif){kind=spacer}
