Kennisbank

AI in hospitality: which work is already changing

A sector of short tasks and a lot of face-to-face contact

Hospitality consists of a large number of small, quickly successive tasks. A kitchen works with orders, mise-en-place, preparation and service. On the floor there are reservations, service, payment and complaint handling. Behind the scenes, there is purchasing, inventory management, staff scheduling and administration. Much of that work is short in duration, but the total number of hours lies mainly in two types of activity: contact with the guest, and organizing what is needed for that. These two types of work respond very differently to what AI can do today.

The moment of contact with the guest is where hospitality businesses distinguish themselves. A waiter who senses the atmosphere, a cook who improvises around an allergy, a host who resolves a difficult situation: this is work in which people make choices based on what they see and hear at that moment. Around that sits a layer of organizing work that is more easily captured in rules and patterns. That layer is where change becomes visible first.

Where AI is already taking over tasks today

Reservations, confirmations and reminders are already handled automatically in many hospitality businesses. A system that accepts bookings, assigns tables and follows up on no-shows does so without an employee reading along, unless an exception arises. Orders placed with suppliers often follow a fixed pattern: consumption over the past period, season, and a minimum stock level. Where that pattern is stable, a system can suggest or automatically place the majority of order lines, with an employee assessing the outliers.

Staff scheduling is a third area that takes up many hours and where automation is already well advanced. A scheduling system that combines historical busyness, weather forecasts and requested leave can produce a first draft roster. The question of which part of that scheduling system AI can handle and which part remains with a manager is worked out on the page about which work AI can take over in scheduling. Recording hours — clocking in, clocking out, breaks, overtime — is a similar case: a closed, repeatable process, described on the page that shows whether AI can take over the recording of worked hours.

Why one hospitality business is further along than another

The shift does not proceed evenly. A chain with multiple locations and a central point-of-sale system already has the data in place to automate orders and scheduling; an independent business with a cash book and a chalkboard for the roster lacks that foundation. The type of concept also plays a role: a fast-food chain with a limited, repeated menu lends itself better to predictable purchasing than a restaurant that varies with what the market offers. And oversight of the exceptions — the allergy, the complaint, the full terrace on an unexpectedly sunny day — remains human work, regardless of how far automation of the routine has progressed.

This mix of repetition and exception is not unique to hospitality. It is reminiscent of the way seasonal work and weather dependency drive automation in the agricultural sector, and also of the balance between system and guest orientation seen in the question of which work in the recreation sector changes first. In both cases, the degree to which the work is predictable determines how quickly AI takes over a task. The administrative side of hospitality — VAT remittance, cash register closing, invoices to suppliers — in turn connects to what is seen in financial services, where administrative tasks have been taken over for longer already. Anyone wondering how the systems that make this possible are built and maintained will find that covered in the work in the ICT sector that is itself also changing due to AI.

What remains, and what shifts within a role

Few roles in hospitality disappear entirely; more often, the composition of the work shifts. A planner spends less time filling in a roster and more time assessing exceptions. A buyer checks proposals instead of assembling every order themselves. A host who used to take reservations by phone focuses on the guests who are already there. That shift concerns the content of the work, not the question of whether someone remains employed. Decisions about personnel fall under their own legal requirements and are not part of this explanation.

What you can do now

This page describes the sector in general terms. To see how this applies to a specific hospitality business, a concrete picture of its own tasks is needed: which tasks there are, how many hours they take and how predictable they are. A free quickscan is available for this: twelve questions, no account required, resulting in an indication of what portion of the hours in that profile can be taken over by AI today. The full work scan, which breaks down a company's work task by task and calculates it into fte capacity per task, is still under development.