AI in hospitality: what an AI-manager does in a small property

For a B&B, a guest house, a small hotel or a few holiday homes, AI in hospitality means an AI-manager. It answers guests at any hour, in the language they wrote in, and adjusts your rates between a lowest and a highest rate you set. It also passes on the reviews and messages that need a person, turns requests into tasks and drafts texts such as a room page. You need no technical staff to set it up or look after it.

Answering guests at any hour, in their language

Guests ask the same questions again and again: what time check-in starts, where to park, whether breakfast is included, how far it is to the beach or into town. They get an answer day or night, in the language they wrote in, and the same answer every time.

What an AI-manager answers, and what it leaves for a person, depends on the house information you give it. Say a guest writes at three in the morning to ask where to park. The reply gives the parking details you entered and goes out while you're asleep. If your house information says nothing about parking, the question goes to a person instead.

A question about dates can turn into a booking in the same conversation. The answer comes from the live calendar that your booking engine reads, so the rooms on offer are the ones still available on those nights.

Some messages need a person: a complaint, a request you can't meet, a question your house information doesn't cover, a guest who is upset. Those go to you or a colleague with the whole thread attached, and the guest is told that a colleague will come back to them.

For each job, you choose whether the AI-manager only drafts a reply or sends it directly, and you can change that choice later. At the start every reply can be a draft that you read, correct where needed and send yourself.

Guests write on WhatsApp, on Instagram, in the messages on Booking.com or Airbnb, and by email. An AI connected to only one of those channels sees part of what your guests send. With every channel in one unified inbox, there is one thread per guest and per booking. A question on WhatsApp and a follow-up by email then sit in the same thread, where your team and the AI-manager both read them.

Adjusting rates with demand

For every room type you set a lowest and a highest rate. An AI-manager raises and lowers the rate with demand and never sets it below the lowest or above the highest. It also takes the local events calendar into account, with the concerts, fairs and marathons in your area.

You can still set any single night by hand, and that rate stays as you set it. Every rate change goes into a log you can read. The post on hotel revenue explains how to choose the two rates for each room type.

Reviews that need a reply, flagged as they come in

An AI-manager reads every new review on Booking.com and Google as it comes in. Those that need a reply are flagged, and so are those about a problem in the property, such as a shower that runs cold.

A reply to a review is public, and it stays under the review where later guests read it too. A guest who wrote out what went wrong should get an answer from the person responsible for putting it right.

From a guest message to a task

A message that asks for something becomes a task, assigned to someone on your team, without anyone retyping it.

Suppose a guest reports late in the evening that the lamp in their room has stopped working. The AI-manager turns the message into a maintenance task for that room, assigned, with the guest's words attached, and the guest gets a reply that evening.

A request for an extra pillow becomes a task in the same way. So does a late check-out, which changes when the cleaner can start on that room.

Each morning you get a briefing of the night: the questions answered, the questions handed over, the rates changed and the tasks created.

Room pages and the house guide, drafted for you to edit

Give an AI-manager the photos and the amenities of a room, and it drafts the room page. It can also draft the first version of a whole website, which you then read page by page.

The house guide is drafted from what guests keep asking in their messages. The guest reads it on their own guest page, with the instructions that apply now.

Every one of these texts is a draft until you have read it. You correct what is wrong about your place and add what only you know, and then it goes live.

What an AI-manager should not do

An AI-manager should not pretend to be a person. Article 50(1) of the EU AI Act, Regulation (EU) 2024/1689 (opens in a new tab), covers AI systems that talk directly to people. Such a system must be built to tell people they are dealing with an AI, unless that is obvious to a reasonably well-informed person. The rule has applied since 2 August 2026, and the duty sits with the provider, the company that develops the system.

At your property, a guest who asks whether a person is reading gets a straight answer. When a reply tells a guest that a colleague will come back to them, someone from your team then answers in that thread.

Some decisions stay with you: the lowest and the highest rate, whether to take a group you're unsure about, and how to answer a complaint about your staff.

Guest messages, booking details and preferences are personal data. They belong in your own booking system, under your control, rather than in a chat service with its own terms. Before you connect any provider, ask where a conversation is stored and who else can read it.

Drimbu AI

Drimbu AI is the AI-manager in Drimbu, and it comes with the Flow AI package. It does each of the jobs above inside the same system as your bookings, the unified inbox and the tasks. A task it creates goes into the task and housekeeping list your team works from. Guest data stays in your own booking system, under your control, and you set your rates in bookings and payments.

How to start with AI in hospitality

1. Connect your channels to one inbox and let the AI-manager draft replies for a few weeks. Read every draft before you send it: you'll see what it handles and where your house information has gaps.
2. Set a lowest and a highest rate per room type, switch on demand-based pricing and read the log every week. Widen or narrow the range from what you see there.
3. Let guest messages become tasks, and read the morning briefing before the day starts.

Every step can be switched back, and you can stay on the first one for as long as you like. This week, write down the questions guests asked you and the answers you gave, because that list is the start of your house information. Then start for free with Drimbu, or first read what the AI-manager does.