Concept buildA worked example of how we would build this for a typical business. Not client work: no client, no client data, no achieved results.
AI Chatbots
Concept buildA concept build for a boutique hotel in the UAE whose front desk answers the same guest questions in several languages, day and night. The concierge replies on WhatsApp and the hotel website in the guest’s language, answers from the hotel’s own guide, turns requests into housekeeping tickets with the room number, and puts complaints in front of the duty manager straight away.
The brief
A typical boutique hotel in Dubai or Abu Dhabi hosts guests who write in Arabic, English, Russian, Hindi and more, at every hour. Most messages are simple, such as pool hours, late checkout, extra towels or a taxi to the airport, but each one interrupts a small front-desk team, and requests passed on by phone or on a note get lost between shifts. Meanwhile the message that matters most, a guest unhappy with the room, looks the same in the queue as a question about breakfast, and is often read too late to fix before checkout.
Explainer
Explainer · 7 slides · about 46 s
Concept buildConcept build: Multilingual Guest Concierge for a Boutique Hotel
Concept build, explained in 7 slides: a concept build for a boutique hotel in the UAE whose front desk answers the same guest questions in several languages, day and night.
Slide 1 of 7
01 · AI chatbots
Multilingual Guest Concierge for a Boutique Hotel
Concept build: Multilingual Guest Concierge for a Boutique Hotel.
02 · The problem
What a typical boutique hotel is up against
A typical boutique hotel in Dubai or Abu Dhabi hosts guests who write in Arabic, English, Russian, Hindi and more.
03 · The approach
How we would build it, in 5 steps
We would build it in 5 steps, from the hotel guide as the single source to pilot on one channel, then widen.
04 · Architecture
How the pieces connect
How the parts connect, from the first input to the person who reviews the result.
05 · Demo · sample run
An Arabic question, a towel request and a complaint
A sample run, step by step, including where a person steps in.
06 · Typical targets
What we would measure it against
Typical targets, each with the basis it would be measured against.
07 · Next step
Scope a build like this for your business
Book a strategy call to scope a build like this.
Architecture
Architecture of the Multilingual Guest Concierge for a Boutique Hotel concept: WhatsApp → LLM + translation (message); Website chat → LLM + translation (message); Hotel guide → LLM + translation (guide section); Booking check → LLM + translation (room verified); LLM + translation → Duty manager (complaint alert); LLM + translation → Housekeeping queue (service ticket).
Diagram with 7 parts: WhatsApp (source), Website chat (source), Hotel guide (source), Booking check (processing), LLM + translation (model), Duty manager (human), Housekeeping queue (system of record).
Connections:
Sample run
Guest asks pool hours in ArabicRoutine
The reply is in Arabic, from the hotel guide, with the guide section shown.
Guest asks for extra towelsRoutine
Room number checked against the booking; a housekeeping ticket is created.
Ticket confirmation sentRoutine
The guest gets the ticket reference; housekeeping sees a structured task.
Guest: “the room is too noisy”Needs care
A complaint is detected, so the concierge stops answering on its own.
Duty manager takes overA person takes over
Alert with the thread and the room; the guest is told a person is on it.
Build plan
We turn the hotel’s guest information, from facilities and hours to policies, dining, transport and nearby essentials, into a structured guide the front office owns and edits. The concierge answers only from that guide and shows which section it used, so a changed pool schedule is corrected once, not in every reply.
The concierge detects the guest’s language and replies in it, with Arabic shown right to left in the web widget. Translation runs through the LLM, but each language is enabled only after a test set of guest-style questions passes review by a fluent speaker. Untested languages get a polite English reply and a staff handoff.
When a guest asks for something, such as towels, a cot or a late checkout, the concierge confirms the request, checks the room number against the guest’s booking and creates a ticket in the housekeeping or PMS ticket queue. The guest gets a ticket confirmation, and staff see one structured task instead of a chat to interpret.
Complaints, safety issues, payment questions and anything the guide does not cover skip the queue: the duty manager is alerted with the thread and the room, and the concierge tells the guest a person is taking over. Tone and keyword rules make this call, and both are reviewed with the front office before launch.
We start with in-house guests on one channel, review conversations daily for the first two weeks and fill guide gaps as they appear. Pre-arrival messages and the second channel come after the pilot holds steady.
Typical targets
Typical target
Replies in the guest’s language, around the clock
Basis: Automated replies on guide-backed questions; languages limited to those that passed testing in the pilot
Typical target
Every service request becomes a ticket with the room number
Basis: Design property: structured ticket creation through the housekeeping or PMS ticket API
Typical target
Complaints reach the duty manager in minutes, not at checkout
Basis: Alert-on-detection design; time from message to manager acknowledgement is measured in the pilot
Typical target
30–50% of guest messages answered without front-desk involvement
Basis: The share of guide-answerable questions in a sample of past guest messages, confirmed in the pilot
Your project
Questions taken from your tickets and transcripts, each paired with the source that answers it, with a pass threshold agreed before launch.
The knowledge index, model accounts and channel numbers are set up under your business, so the data and the configuration stay yours.
Message templates, consent wording and widget placement checked against each channel’s rules where your customers are.
Handoff and resolution tracked per intent; scope widens only when the numbers hold for a full week.
Where this fits
FAQ
No hotel runs it. This page is a concept build describing how we would set up a multilingual concierge for a typical boutique property in the UAE. The Arabic exchange in the demo is sample data, and each outcome listed is a typical target with the basis we would check during a pilot.
Any language the model handles well can be switched on, but only after it passes a test set reviewed by a fluent speaker. A typical first version covers Arabic and English; further languages follow one at a time, chosen from what guests actually write in.
Only narrow access: it confirms that a room number matches an active booking and it creates service tickets. It reads no payment details and no more of the guest profile than a ticket needs, and the hotel can revoke its API key at any time.
Not in the first version. It answers questions about them from the hotel guide and turns a booking request into a ticket for the concierge desk, which checks availability and confirms. Direct booking would need write access to each outlet’s system and a payment step, so it is a later decision for the hotel rather than a default.
It acknowledges the guest at once, says when someone will follow up, and alerts the staff member on night duty. Safety-related messages also trigger the hotel’s existing night escalation, such as a call to the duty manager, so nothing urgent waits for the morning shift.
Worked examples
Concept builds are worked examples, not client work.
Concept buildAI Chatbots
A concept build showing how we would build a support bot for an online electronics store whose helpdesk fills with the same few questions. The bot would answer order, delivery, returns and product-fit questions from the store’s catalogue, policies and live order data, on the website and WhatsApp, and pass everything else to the helpdesk with the conversation attached.
Concept buildAI Chatbots
A concept build for a direct-to-consumer brand in India whose WhatsApp inbox fills with the same two questions: where is my order, and how do I return it. The assistant answers in English, Hindi or Hinglish from live order data and the brand’s own returns policy, and hands anything unusual to a person with the whole thread attached.
Concept buildAI Chatbots
A concept build for a B2B software company whose help centre is thorough but hard to search. The copilot answers product questions from the published documentation, shows the passages each answer came from, and opens a support ticket, carrying the question and the sources it tried, whenever it is not confident.
Next step
Book a free 30-minute strategy call. Bring one process you want to automate and we will talk through scope, integrations and a sensible first step.
Free 30-minute call · no commitment
