Reading and categorising enquiries
An email with travel dates and wishes is easy for a language model to understand. 40 enquiries a week become 40 prepared offers. The win is in response time, not in headcount.
Knowledge
AI in the hotel:
what it can do and where it has limits.
Many hotel chains are already using AI. What is often missing is a shared direction, clear responsibilities and clean integrations. That is where isolated tools, unnecessary risks and team frustration appear.1
1 The global h2c study shows: 78 percent of surveyed hotel chains use AI, but only 7 percent have a company-wide AI strategy. 189 people from 171 hotel chains worldwide were surveyed. Source: h2c, AI & Automation in Hospitality, 2025. Read the study
Where AI in the hotel
really helps today.
In six places very clearly. What they share: a lot of repetition, clear rules, a verifiable result.
Texts in the guest's language
Confirmations, reminders and replies in six languages, in the house's tone. What used to be a cost centre is now a setting.
Finding patterns in numbers
Occupancy curves, cancellation rates, source markets. AI spots anomalies in data you already have, before they show up in the monthly report.
Triggering recurring workflows
Room status, task allocation, official reporting. Wherever someone works through a list today, a system can make the start and leave the decisions to people.
Making in-house knowledge available
House rules, packages, agreements with regulars. New staff find answers without interrupting someone who is with a guest.
Checking instead of searching
An arrival check that flags what is missing before the shift starts: registration forms, room allocation, bed configuration for families. A small thing, every day.
Where AI in the hotel
still falls short today.
You rarely hear these six points from vendors, but they decide whether a project holds up.
Bad data gets worse faster
AI does not turn a messy dataset into a clean one. It just distributes it faster. Inconsistent room categories and duplicate guest profiles become serial errors instead of being caught.
The last step stays manual
A system can write an offer. Whether this regular guest gets a discount this year is still a human call. Forcing automation where judgement is needed loses guests.
Confidently wrong
Language models answer smoothly even when the ground is thin. Without a visible data snapshot and a confidence score, nobody in operations notices. Ask every vendor how they solve this.
Costs that grow with you
Every model request costs money. Without a quota, a usage display and a hard cap, a predictable line item turns into an open invoice. The most common nasty letter after the first quarter.
Team acceptance
A receptionist who feels replaced will not use the tool. One who sees the tedious twenty morning minutes disappear, will. That is a change-management issue, not a technical one.
Vendor lock-in
If the automation rules only live in the vendor's system and cannot be exported, a later switch is expensive. Ask about export formats and timelines before signing.
What AI-native
really means.
The term is now on almost every vendor's page. It has, however, a testable meaning.
AI-native means: AI is the way the software is operated. It is not an add-on beside the actual product. Data model, permissions, protocol and interface are built so that the system prepares and the human decides.
The counterpart is called bolted on in the trade: an existing piece of software with a chat window added on top. There, the chat talks to the user, but not to the system.
The simplest test
Imagine every AI function taken away. What is left?
If the product looks the same as before, AI was bolted on. If the operation shifts to where menus and clicks used to be, it is AI-native. The test costs nothing and shows more than any sales pitch.
Seven signs
that AI has only been bolted on.
Take this list into your next vendor meeting. You do not need IT skills for it.
- The AI lives in a drawer. There is a button, a window pops up, the assistant is inside. Everything else you still use as before.
- The assistant only knows one area. It can talk about reservations, but not about housekeeping. It sits on a single interface instead of the shared dataset.
- It talks, but it does nothing. It answers questions and suggests text. As soon as something needs to change in the system, it tells you where to click.
- No data snapshot, no confidence score. You see the answer, but not what it rests on or how sure it is. Software built for AI from day one shows both, because it needs them for its own checks.
- Approvals are not configurable. Either a step runs automatically or not at all. The in-between state does not exist for the system.
- AI usage does not appear on the invoice. No quota, no usage display, no cap. This usually means nobody at the vendor is responsible for model costs.
- Pricing hangs solely on the number of users. A price per seat assumes only people use the software. That no longer holds for a system that works on its own.
What the law
asks of you.
The EU AI Act assigns duties depending on your role. For hotels, the main ones are a clearly labelled guest chat, trained staff and an auditable use of AI.
Disclosure obligation
Whenever a guest communicates with a chatbot, a digital concierge or an AI-generated reply, they must be able to recognise this. The transparency obligations apply in full from 2 August 2026. Breaches are subject to sanctions.
AI literacy in the team
Businesses that deploy AI must ensure their staff have sufficient AI competence. A short, documented briefing therefore belongs to onboarding.
You are responsible
For guest data you are the controller, not the vendor. You need a data processing agreement, and for high-risk processing a data protection impact assessment.
2 European Commission, EU AI Act: transparency obligations under Article 50 apply from 02.08.2026. The obligation to ensure sufficient AI literacy under Article 4 applies from 02.02.2025. Which duties apply to your hotel depends on the system and your role. European Commission guidance
This page is not a substitute for legal advice. It names the obligations most often overlooked in practice.
Where felixUnity
stands on this list.
The same seven questions, applied to us. You are free to verify each of them in the demo.
| The question | Our answer |
|---|---|
| Does the AI live in a drawer? | felix is how you operate the system, not a window next to it. You write or speak in plain sentences and things happen in the system. The classic views stay alongside. |
| Does it only know one area? | Reservations, offers, housekeeping, official reporting and analytics sit on one dataset. Plus over 400 integrations from 46 years of felix. |
| Does it only talk? | It creates offers, distributes tasks, generates lists and notifications. What it may execute is up to you. |
| Does it show a data snapshot and confidence? | Yes. Every check shows the timestamp, the data snapshot and the areas covered; incoming enquiries carry a confidence score. |
| Are approvals configurable? | Yes, per workflow and per role. You move the boundary in either direction at any time. |
| Is AI usage visible? | Monthly quota, usage visible before triggering, hard cap as default, warning at 70, 90 and 100 percent. |
| And the guest data? | Storage and processing on European servers only. Model training happens without your guest data. |
Where we do not convince today
felixUnity is launching together with selected businesses. We do not yet have reliable impact numbers from live operations. We will publish them once they exist. And not before.
All test questions in the comparisonFrequently asked questions
about AI in the hotel.
Does AI replace people at reception?
In practice, it shifts the work rather than replacing it. What disappears is the typing, looking up and forwarding. What remains is the part guests come back for.
Do I need clean data before I start?
Not across the whole property, but in the starting area. A sensible project begins with a single workflow, so the dataset can be brought into shape in a manageable time.
Do I have to tell guests that an AI is replying?
Yes. Since 2 August 2026, the transparency obligations of the EU AI Act apply in full, also to chatbots and AI-supported communication. The disclosure must be recognisable for the guest.
Is our guest data used for training?
Not at felixUnity. Storage and processing take place on European servers only, and model training happens without your guest data. Ask every vendor to confirm this in writing.
What does AI usage cost per month?
What matters is the structure: a monthly quota, visible usage and a hard cap as the default. Without these three things, planning is impossible.
Put us to the test
on one of your own workflows.
Bring a real enquiry from your hotel. We go through it together, and you ask the seven questions from this page.