How it works

Product

Values

Blog

Register your interest

How it works

Product

Values

Blog

Register your interest

AI

AI in Hotels: Use Cases for Guests, Management, Marketing & Team

AI in hotels explained: how hotels use AI in guest communication, management, marketing and teams, what studies show and which field to start with.

Eric Rosenkranz

Person mit einem Tablet an einem Tisch

AI in hotels is not one tool. It spreads across four fields of work: guest communication, management, marketing and the team. In each, something different is at stake. A person reviews a text draft before it reaches the guest. A chatbot's answer reaches the guest at once. This guide walks through the four fields one by one and shows where starting carries the least risk.

What is AI in a hotel?

AI in a hotel is software that uses machine learning to process text, data or speech and supports tasks in hotel operations, such as guest communication, pricing, review analysis or internal look-ups. Hotels meet it as a stand-alone tool or as a feature inside systems they already use.

The word "feature" matters. The survey this guide draws on notes that hotels probably already use AI features without knowing it, because AI is increasingly built into software. So anyone introducing AI in a hotel should start by taking stock of the systems already running in the house.

Who is this guide for?

It is written for owners and general managers who want to judge in which part of their hotel AI can sensibly begin. It assumes no technical background.

Where do hotels use AI?

The most widespread use is generative AI for text in guest communication. Dynamic pricing, forecasts, review analysis and chatbots follow. Use for employees is the least covered in the available surveys. The data comes from a survey by Roland Schegg (HES-SO Valais-Wallis) with EHL, run from January to April 2025 through hotel associations in Austria, Germany, France, Greece, Italy and Switzerland. 82% of the 1,485 hotels were independent, and the median was 31 rooms.

This guide sorts the reported uses into four fields:

  • Guest communication: texts and messages with generative AI (80%), chatbots (37%), personalised emails and recommendations.

  • Management: dynamic pricing (50%), forecasts (44%), workforce planning.

  • Marketing: review analysis (44%), replies to guest comments, texts and images.

  • Team: knowledge, onboarding, training. The report gives no figure for this in DACH. Workforce planning was named by 15% of hotels using AI across all six countries.

All shares refer to hotels that already use AI (41% of respondents) and apply to DACH in 2025 unless stated otherwise (voluntary survey, not representative).


Six dot grids: share of DACH hotels using AI per application in 2025, from 80 of 100 for texts to 2 of 100 for facial recognition

How has use developed?

Across all six countries, the survey reports 41% current AI use, 43% no use and 16% planned adoption. In DACH, among hotels using AI, the share of generative text tools rose from 19% to 80% between 2023 and 2025, chatbots from 17% to 37% and dynamic pricing from 34% to 50% (the 2025 sample includes smaller hotels, so the comparison is only a rough guide).

76% of hotels using AI named time savings as the biggest benefit, followed by better communication and marketing (54%) and higher operational efficiency (51%). These are self-assessments, and impact was not measured. As barriers, 39% of respondents named little knowledge of available solutions, 35% high setup costs and 34% technical complexity.

AI in guest communication: where does it help, and where does it need sign-off?

In guest communication, hotels use AI mostly for text drafts and less often for chatbots that write directly to guests. The difference is decisive: a person reviews a draft before it reaches the guest. A chatbot's answer reaches the guest immediately and, in the guest's eyes, speaks for the hotel.

What hotels use here

80% of DACH hotels using AI use generative text tools for guest communication, up from 19% in 2023. Chatbots are used by 37% (2023: 17%). Across all six countries, 38% use personalised emails, recommendations or guest apps and 31% use chatbots. In the open answers, hotels also named AI-supported phone assistants, machine translation for multilingual operations and planned interpreting services at reception.

Which tasks fit

The following list is an editorial compilation of tasks where hotels can use AI in guest communication:

  • Draft replies to enquiries, for example about availability, pets or parking. A team member reviews and sends.

  • Pre-arrival messages with directions and check-in time, in the guest's language.

  • Translation of incoming and outgoing messages.

  • A chatbot for recurring questions on the website or in a messenger, with a clear handover to staff.

  • Phone cover outside staffed hours. Hotels in the survey named night shifts and low-staff periods as a possible area for self-service.

What studies show

There is little research so far on how chatbots work in hotels (a 2023 review covered 48 papers). In the survey's open answers, hotels fear a loss of personal contact, and some want to stay with analogue processes on purpose.

Why sign-off matters is shown by a case outside the hotel industry: in 2024 a Canadian tribunal ruled that an airline was liable for wrong information its website chatbot gave about bereavement fares. The tribunal treated the chatbot as part of the website, for whose content the company is responsible (not a hotel case, not binding precedent).

What matters

  • Draft or direct answer? Decide which messages a person reviews before they reach the guest.

  • Binding information on prices, cancellation terms, accessibility and safety needs an approved source. Without one, it stays with staff.

  • A way to say "I don't know." An assistant with no limits invents answers.

  • Spot checks with real questions. Test with the enquiries of the last few weeks, not ones you made up.

How to start

Collect 30 guest enquiries from recent weeks and sort them into three piles: questions with always the same answer, questions that need judgement, and complaints. Start with drafts for the first pile. People keep answering the third.

AI in management: pricing, forecasts and planning

In management, AI supports pricing, occupancy forecasts and planning. It estimates demand from past bookings and market data. The decision stays with the general manager. In the survey, 50% of DACH hotels using AI use dynamic pricing and 44% use forecasts.

What hotels use here

Among DACH hotels using AI, 50% used dynamic pricing in 2025 (2023: 34%) and 44% used forecasts ("predictive analytics"; 2023: 37%). Workforce planning was used by 15% of hotels using AI across all six countries in 2025, against 29% in 2023. The report puts the drop down to more small hotels in the sample. 17 to 20% of respondents plan to introduce forecasts and automated replies.

Which tasks fit

Editorial compilation:

  • Price recommendations per night and room category, which you approve or adjust.

  • Occupancy forecasts that make a weak week visible early.

  • Staffing needs derived from the forecast and used as a draft for the rota.

  • Monthly reports, where AI turns figures into a first draft text.

What studies show

Forecasting has the strongest evidence. A study using daily data from a four-star hotel in southern Europe compared 22 forecasting methods. Machine learning methods cut forecast error (RMSE) by up to 54% at a one-day horizon and by up to 45% at 14 days, compared with classic exponential smoothing (data from a single hotel). A 2024 review of 20 papers sees potential for more accurate forecasts and names the effect on pricing and financial results, data quality and expertise as open points.

We found no independent study on the effect of AI pricing on hotel revenue.

What matters

  • Data. A forecast is only as good as the booking history. Missing cancellations, unmaintained room categories or one-off large groups distort it. In the survey, larger hotels name poor data quality and integration as a barrier.

  • Set limits. Define minimum and maximum prices before a system suggests any.

  • A person approves, especially for events, trade fairs and exceptions the history does not know.

How to start

First check whether your property management system or channel manager already includes price recommendations or forecasts. For four weeks, compare its suggestion with your own decision without adopting it. You will see where suggestion and experience differ.

AI in marketing: copy, images and reviews

In marketing, hotels use AI for text and images, for analysing guest reviews and for replying to them. Results stay drafts at first, which makes this the easiest place to start. 44% of DACH hotels using AI use review analysis, and 62% of all respondents see marketing content as a promising field.

What hotels use here

In DACH in 2025, 44% of hotels using AI used review analysis (2023: 38%). Across all six countries, 37% automate replies to guest comments and 19% generate images for content (2023: 3%). 62% of all respondents named marketing content as one of the most promising fields for AI, behind reservations at 68%. In the open answers, hotels mention texts for newsletters and social media, translation, replies to reviews and detecting fake negative reviews.

Which tasks fit

Editorial compilation:

  • Pull themes out of reviews: which points recur in the last 100 reviews, such as breakfast, noise or check-in?

  • Draft replies to reviews, which a person adapts and publishes.

  • Texts for website, newsletter and social media based on your facts.

  • Descriptions for packages and seasonal offers.

  • Translation of content into your guests' languages.

What studies show

Replying to reviews works, as a study of hotels in Texas shows (replies written by people, before generative AI). Hotels that started replying on TripAdvisor received on average 0.12 stars more and 12% more reviews, compared with Expedia, where hardly anyone replies. The authors do not put this down to better quality but to who decides to write a review: dissatisfied guests write fewer short, unsubstantiated bad reviews when the hotel replies.

A small study in Cornell Hospitality Quarterly showed participants real reviews with a hotel's reply and a ChatGPT reply. ChatGPT imitated real replies fairly well. The authors advise considering perceived authenticity before hotels hand the task to AI.

What matters

  • The facts come from you. AI words them, you supply room numbers, opening hours, facilities and prices. A text must not promise what the house does not deliver.

  • Make review replies personal. Pick up details from the review and sign with a name.

  • Do not paste guest names into unchecked services. For public reviews, the text without the name is enough.

How to start

Copy the last 100 public reviews without names into a tool of your choice and have it name the three most frequent themes. Read 10 reviews yourself to check whether the themes are right.

AI in the team: knowledge, onboarding and scheduling

For employees, AI starts where knowledge is looked up, explained or passed on: answering questions about procedures, supporting onboarding, translating texts, preparing rotas. This field is the least covered in the surveys.

What hotels use here

Workforce planning was used by 15% of hotels using AI across all six countries in 2025. In the open answers, hotels mention training videos with avatars, AI-supported competitor analysis and AI for night shifts and low-staff periods. One respondent names knowledge building as a less tangible but meaningful benefit. As barriers in the team, small hotels name little knowledge of solutions and a lack of technical skills, while resistance to change is stronger in hotels with 18 to 35 employees.

Which tasks fit

Editorial compilation:

  • Answer questions about procedures, for example house standards, allergens or emergency rules, with a reference to the underlying source. This requires maintained hotel knowledge, as our article on hotel knowledge management describes.

  • Draft onboarding plans from existing checklists.

  • Summarise shift handovers when the entries are complete. What belongs in an entry is covered in our article on the hotel shift handover.

  • Translate content for a multilingual team.

  • Draft rotas from the forecast and availability.

What studies show

We found no hotel-specific studies on how AI affects employees' work. Two studies from other industries show a pattern for knowledge work.

A study of 5,172 customer support agents at a software company found that an AI assistant raised the number of issues resolved per hour by 15% on average, with large differences between workers. A field experiment with 758 management consultants found: on 18 tasks within the AI's capabilities, participants with AI access completed 12.2% more tasks, 25.1% faster and at higher quality. On a task outside those capabilities, they were 19% less likely to get it right than the group without AI.

Neither study is from hotels. The pattern: AI helps employees with tasks for which it has good templates and can mislead them on tasks for which it has none. For operations, this argues for tying answers to maintained, approved sources.

What matters

  • Knowledge first. An assistant can only answer as well as the knowledge it draws on. Outdated standards produce outdated answers.

  • Involve the team. Let employees help decide which questions the assistant should answer.

  • Rotas and personnel data need extra care, because they concern employees.

How to start

For one week, note which questions your team asks colleagues or managers most often. For the first ten, check whether a current written answer exists. Where it does not, knowledge is missing first, not the tool.

What does this look like day to day? An invented example

The following example is invented. It shows where, in a 30-room hotel, a person decides each time. It is not a customer story and says nothing about results.

Monday, 8 a.m., front desk:

  • Guest communication. An enquiry asks whether dogs are allowed. AI drafts the reply from the house rules. The receptionist checks the price for the dog bed and sends it.

  • Management. The forecast shows a weak week in three weeks. The general manager sees the suggestion, adjusts the price for two room categories and approves it.

  • Marketing. Three recent reviews say breakfast ends too early. AI summarises the theme and drafts a reply. The restaurant manager adds that hours are longer from May and publishes it.

  • Team. A new temp asks what to do about a nut allergy. The assistant answers with the approved procedure and names the source. If the procedure is missing, it points to the kitchen manager.

In all four cases, AI delivers a draft, a summary or an answer from a source. A person makes the decision each time.

Which field should you start with?

Start with a process, not a tool. The sources show no generally best starting point and no ranking by benefit. For each possible use, check four things:

  1. What data goes in? Without personal data, starting is easier.

  2. What happens if it goes wrong? A wrong draft that a person reviews weighs differently from a wrong answer to a guest.

  3. Is AI already in your systems? Check your existing contracts.

  4. How will you measure whether it helps? Response time, correction rate, escalation rate and effort in the team are sensible measures for a test. They are not documented as AI results.

From these questions we derive a recommendation. It is editorial, not a study result:

  • Marketing has the lowest cost of error, because drafts are reviewed and guest data is often not needed.

  • Guest communication pays off when the same questions keep coming back. Start with drafts, not a chatbot.

  • Management pays off when your booking data is maintained and your system already offers suggestions.

  • Team pays off when important knowledge sits with individual people. It needs maintained knowledge first.


Two-by-two map of the four AI fields by cost of error and preparation needed: guest communication and management with higher costs of error, marketing and team with lower, marketing as the recommended start

Which rules should you know?

Two sets of rules affect the use of AI in hotels: the EU AI Act and the GDPR. Article 4 of the AI Act has required operators since 2 February 2025 to take measures to support the AI literacy of their staff. Since the amendment of July 2026, the text states that no specific level of literacy has to be guaranteed, and national market surveillance authorities have enforced the rule since 2 August 2026. The transparency duties of Article 50, under which people must be told they are talking to AI, have applied since 2 August 2026. The GDPR applies as soon as personal data enters an AI system. We cover rules and risks in detail in a separate article.

Which mistakes should hotels avoid?

  • Pasting guest data into unchecked public services.

  • Giving out binding information unchecked, especially on prices, cancellation and safety.

  • Concluding benefit from usage figures. That many hotels use text tools does not prove they pay off.

  • Starting with the tool instead of the process.

  • Leaving the team out.

How Vesca approaches AI

Vesca starts in the Team field: a hotel assistant built around how your hotel works answers employees' questions about procedures and standards and shows the source. If information is missing, the assistant says so, and the hotel can add it. Managers confirm what is current and stay in control of the approved knowledge. The assistant supports your team and does not replace it. Discover more digital use cases for hotels at Vesca. Vesca is in development and will be available to hotels soon.

Get involved

If you would like to help shape Vesca with your hotel, you can register your interest without obligation. We select suitable hotels and plan a free pilot project with each of them. Registering is not a pilot commitment and not a purchase commitment.

Register your interest

Read next

Sources

Author

Eric Rosenkranz

Eric Rosenkranz is the founder of Vesca Systems GmbH. Vesca is being built with independent hotels in DACH.

Last reviewed:

‹ Hotel shift handover: process, content & digital shift log

Hotel Onboarding: Process, Checklist & Mistakes to Avoid ›

Hotel Onboarding: Process, Checklist & Mistakes to Avoid ›