Artificial intelligence in hospitality: uses, guest data and the EU rules
Artificial intelligence in hospitality is less about robots in the lobby than about the decisions hotels, restaurants and holiday rental operators make every day: what to charge tonight, how many people to roster on Saturday, how to answer the same guest question for the hundredth time, which rooms to clean first. For artificial intelligence in hotels, most of the value sits in the property management system, the channel manager and the guest inbox rather than in new hardware. This guide covers where AI helps in hospitality, the data it needs, the EU rules on guest and staff data, and how to start with one use case that pays back.
Where AI is used in hospitality
| Area | What AI does | Data it needs |
|---|---|---|
| Pricing and revenue | Suggests room or package rates from demand, pace of bookings, events and competitor rates | Booking history, availability, rate history, local events |
| Demand forecasting | Predicts occupancy, covers and arrivals by day and hour, which drives staffing and purchasing | Bookings, sales, weather, calendar |
| Guest messaging | Answers pre-arrival and stay questions in the guest's language, takes simple requests, hands the rest to staff | Your policies, room and service information, booking data |
| Reviews and feedback | Groups comments by topic and mood across platforms, drafts replies for staff to approve | Reviews, surveys, complaint logs |
| Housekeeping | Orders rooms by departures, arrivals and priority guests, balances workload across the team | Room status, booking data, staff rosters |
| Maintenance and energy | Heats, cools and ventilates rooms by occupancy, flags equipment that is drifting before it fails | Room sensors, meters, building systems, bookings |
| Kitchen | Forecasts dishes and covers to cut food waste and purchasing errors | Sales by dish, bookings, menu changes |
| Back office | Reads supplier invoices, matches them to orders, prepares reports | Invoices, purchase orders, accounting data |
Text and language are where most businesses start, in hospitality as elsewhere. Eurostat reported on 11 December 2025 (read 11 November 2026) that 20.0% of EU enterprises with 10 or more employees used AI technologies in 2025, and that analysing written language was the most common use, at 11.8%. For a hotel that means guest messages, reviews and supplier documents.
Pricing and forecasting: the highest value, the most care
Revenue management has used forecasting models for decades; what AI adds is more signals, more frequent updates and forecasts for smaller properties that never had a revenue manager. A good pricing system shows its reasons (pace of bookings, an event in town, a competitor sold out) and keeps the rules the business sets: minimum and maximum rates, parity with distribution contracts, length-of-stay restrictions. The forecast that drives the price should also drive the roster and the purchasing, so the hotel is not priced for a full house and staffed for a quiet night.
Prices shown to guests remain subject to consumer law, and charging individual guests different prices based on their profile is treated differently from changing prices with demand. Our AI dynamic pricing page covers the EU price rules in detail.
Guest messaging and chatbots
Most guest questions repeat: parking, check-in time, breakfast hours, late checkout, how to reach the hotel from the airport. A messaging assistant that answers from the property's own information, in the guest's language and at any hour, takes that load off the front desk. Done well, it also takes simple requests, such as extra towels or a table booking, and passes them to the right team with the room number.
Three rules keep it trustworthy. It answers only from approved content and says when it does not know. It hands over to a person when the guest asks or the topic is a complaint, a refund or a safety matter. And it says it is an AI: the EU AI Act, Regulation (EU) 2024/1689, Article 50(1), requires that people be informed they are interacting with an AI system unless that is obvious. See chatbot development for how we build answers from your own content.
Guest data is personal data, and sometimes sensitive
Guest profiles, booking histories, preferences and messages are personal data under the GDPR, Regulation (EU) 2016/679. Some of what hotels and restaurants learn goes further: dietary needs and allergies can reveal health or religion, accessibility requests reveal health, and a face used to identify a guest is biometric data. Article 9 treats those as special categories, which may be processed only on specific grounds such as explicit consent. In practice:
- collect what the stay needs, and keep sensitive notes out of general marketing profiles;
- tell guests what AI does with their messages and data, in the privacy notice and at the point of use;
- check what the AI provider may do with the data, and keep data in an EU region where possible;
- set retention: guest messages and profiles are not kept forever by default.
Holiday rental operators have a further set of rules: Regulation (EU) 2024/1028 on short-term accommodation rental data, which applies from 20 May 2026, sets how online platforms and hosts collect and share registration and activity data with public authorities (EUR-Lex, read 11 November 2026). A system that holds that data should be built with the reporting in mind.
Staff data and the AI Act
Some hospitality AI touches employees rather than guests, and that is where the AI Act is strictest. Using AI to infer the emotions of people in the workplace is prohibited under Article 5(1)(f), outside medical or safety reasons, which rules out tools that read staff moods from cameras or voice. AI used to allocate tasks based on individual behaviour or personal traits, or to monitor and evaluate the performance of workers, is listed as high-risk in Annex III, point 4(b), with the main duties applying from 2 December 2027. A housekeeping planner that orders rooms by departures and arrivals is a different thing from one that scores each housekeeper; the second needs the high-risk process, human oversight and worker information. We draw that line at the start of a project. For rosters themselves, see employee scheduling software.
The AI Act's general application date is 2 August 2026. Biometric identification at a distance is high-risk under Annex III, point 1(a), while pure verification that a person is who they claim to be is excluded from that point, though it remains special-category data under the GDPR.
Connecting to the systems a hotel already runs
Hospitality AI is only as good as its access to the property management system, the channel manager, the booking engine, the point of sale and the building systems. Most modern property management systems, such as Oracle OPERA Cloud, Mews and Apaleo, offer interfaces for exactly this, while older on-premise systems may need exports or a middleware layer. Integration is usually the largest part of the work and the part that decides whether staff trust the output, because a forecast built on half the bookings is wrong in ways nobody can see. Our API integration page explains how we connect systems with retries, logs and data checks.
How to start an AI project in a hotel or restaurant group
- Pick one recurring decision or task. Weekend staffing, the first answer to guest messages, rate suggestions for one property, food purchasing for one kitchen.
- Measure how it is done today. Hours spent, error rates, waste, response times, so the result can be compared honestly.
- Check data and rules. What exists in which system, whether it includes sensitive guest or staff data, and which AI Act category the use falls into.
- Run a small version with the people who own the task. Staff approve suggestions before anything is automatic.
- Compare with the baseline after a full cycle. At least a few weeks, longer for anything seasonal, then extend, adjust or stop.
For the general method, from task to tested program, see how to make an artificial intelligence program.
What we build for hospitality
We build the software around the models as well as the models: guest apps and messaging assistants, forecasting that feeds rosters and purchasing, dashboards that combine property, sales and review data, integrations with property management and point of sale systems, and staff apps. Data stays in an EU region in the client's own account, and the code belongs to the client. See AI consulting for choosing a first use case, restaurant app development and travel app development for guest-facing apps, and data analytics consulting for reporting.
Questions people ask
How is artificial intelligence used in hospitality?
Mainly for room and table pricing, demand forecasts that drive staffing and purchasing, guest messaging in several languages, reading reviews and feedback at scale, housekeeping and maintenance planning, energy control by occupancy, and food waste forecasting in kitchens. Most of these work behind the scenes and support staff decisions rather than replacing guest contact.
Will AI replace hotel and restaurant staff?
Not in the uses that work today. The common pattern is that AI takes the repetitive part, such as answering the same pre-arrival questions, drafting replies or proposing a rate, and staff handle the exceptions and the guest contact. Hospitality is judged on service, and the businesses that use AI well use it to give staff more time with guests.
What data does a hotel need for AI?
Booking history from the property management system, rates and availability from the channel manager, sales from the point of sale, guest messages and reviews, and for energy uses, room sensor and meter data. Two or more years of booking history make demand forecasts far more useful, because they include the full seasonal pattern.
Do hotel chatbots have to say they are AI?
Under the EU AI Act, Article 50(1), providers must design AI systems that interact directly with people so that people are told they are interacting with an AI system, unless that is obvious from the context. A guest messaging assistant should say so, and should hand over to a person when asked or when it cannot help.
Is face recognition at hotel check-in allowed?
Using a face to confirm that a guest is who they claim to be is biometric verification. The AI Act excludes pure verification from its high-risk biometric list, but under the GDPR biometric data used to identify a person is special-category data, which in practice means explicit consent and a real alternative, such as a passport check at the desk.
Sources
- Regulation (EU) 2024/1689, AI Act, Articles 5(1)(f), 50(1) and 113, Annex III points 1 and 4 (read 11 November 2026)
- Regulation (EU) 2016/679, GDPR, Article 9 (read 11 November 2026)
- Regulation (EU) 2024/1028 on data collection and sharing relating to short-term accommodation rental services (read 11 November 2026)
- Eurostat, 20% of EU enterprises use AI technologies, 11 December 2025 (read 11 November 2026)