A hospitality business runs on a combination of predictable work and work that only arises in the moment itself. Making schedules, tracking purchasing, counting stock, processing reservations, checking invoices: that is work that repeats itself and whose outcome is largely fixed in rules and historical patterns. Alongside that is the work that happens the moment a guest walks in, a kitchen is under pressure, or a group without a reservation is at the door. That second part often determines the experience on which a guest judges a business, and the first part determines whether there is still any margin left at the end of the month.
The circumstances that drive the outcome are well known: occupancy that varies by part of the day and by season, staff who must be scheduled at short notice, suppliers with variable availability, and a guest who can submit a reservation, order, or complaint at any time of day. Those who win here have so far done so mainly through people who had overview and could switch quickly.
Schedule planning, reservation processing, and the initial handling of inquiries are tasks that lend themselves to being taken over by AI: they are repetitive, there is a lot of historical data, and the decision rules can be made explicit. Where that happens, what a guest or supplier experiences as distinctive shifts. Response speed to a booking used to be a matter of who happened to be on the phone at that moment; if a system answers that question instantly, speed is no longer a distinguishing factor, because everyone who has the system is equally fast.
What remains distinctive, then, is the work where AI does not take over the task but supports it with a human who approves or rejects: a staff schedule that is proposed by a system but adjusted by a manager based on who can handle a difficult group, or a menu change that is proposed based on sales data but judged by a chef on taste and season. And what remains entirely human work is the interaction at the table, the way a complaint is resolved, the atmosphere that staff create on the floor. That cannot be automated because its value arises in the moment itself, not in a pattern that can be derived from data.
The speed at which this changes differs greatly per business. A chain with many locations and a lot of transaction data has a basis for having schedules and purchasing proposed by a system; an independent establishment with a fixed core of regular guests often does not have that basis and also has less use for it, because the value there already lies in the personal relationship. The difference, then, is not in ambition, but in what repeatable work and data is available to build on.
If delivery time, response speed, or availability is increasingly determined by systems, that is no longer a distinguishing claim once competitors use the same systems. A business that still claims to win on the speed of reservation confirmation, while that speed is now the same everywhere, is in fact not winning on that dimension but still thinks it is. That is precisely where comparison with the peer group is valuable: not to confirm what a business already thinks, but to show which dimension still distinguishes and which no longer does.
That same dynamic plays out in other sectors in its own way, as can be seen in the shift in the agricultural sector where AI takes over work and in the consequences for the recreation industry as reservation and planning work is taken over. In the ICT sector, where a lot of this kind of automation has existed for longer, it is easy to see which dimensions there were the first to lose their distinguishing power, found at the analysis of winning dimensions in the ICT sector now that AI takes over work.
The underlying question, which work in a specific hospitality business can truly be taken over by AI and which work remains human work, cannot be answered in general terms. That differs per business, depending on the data available, the systems already running, and the nature of the work itself. The work scan from FTE TO AI answers that question per task, without assumptions about the outcome.
An additional complication in hospitality is that competitors rarely publish figures on occupancy, margins, or turnaround times. Anyone who wants to know on what a competitor actually scores better must work with what is visible: reviews, waiting times, staff turnover, menu changes. How to read those indirect signals is covered in the approach for assessing competitors who do not publish figures. Another problem is that marketing texts in the sector often resemble one another: everyone promises quality, hospitality, and experience, which makes the distinction in words useless to rely on, as elaborated in the analysis of why all competitors sound the same in their promises.
What an HR manager does with this shift in terms of staffing and workforce composition falls under the applicable legal requirements for that; that is not a subject on which a statement is made here.
The first step is not an investment, but a check: which claims about speed, offering, or service are currently used to win, and are they still distinguishing or now the same everywhere. The free dimension check makes that concrete: you name what you think you win on and see which of those claims can be defended with evidence. The full benchmark, with the peer group on all relevant dimensions and an evidence matrix per score, is under construction.