Hospitality is work that concentrates around peaks: the seven o'clock dinner, Saturday night, the weekend with a full calendar of reservations and events. Many hours go to guest contact, to preparation, to service in the moment itself, and to the organization around it — purchasing, planning, rosters, stock. That makes the sector sensitive to two things that seem to have little to do with each other, but both determine what AI can do here: the unpredictability of demand, and the extent to which the guest experience revolves around one person responding to another.
That combination steers the outcome. Work that is predictable and detached from direct guest contact — roster concepts, purchasing proposals, stock signals, staffing planning based on expected busyness — lends itself to being taken over by AI. Work that revolves around the moment itself, around improvisation at the table or around assessing a difficult situation with a guest, remains human work, or at the very least requires continuous human oversight.
In parts of the hospitality sector, the shift is already happening. Reservation systems that themselves propose how best to plan an evening. Purchasing proposals that read along with consumption and spoilage. Roster concepts that take expected busyness into account, which is already a very different task from approving the roster itself or adjusting it to who is sick that week. That last part — the approval, the correction based on something the system could not have known — is the second category: AI makes the proposal, a human approves or rejects it and says why.
The third category, guest contact itself, shifts the least. A guest with a complaint about the service, a team that has to improvise when the kitchen is behind schedule, a sommelier who senses what suits a party — that remains with people, not because AI could not attempt it, but because the value of the contact lies largely in the human element.
Why does this move faster at one company than at another? Often the difference lies not in the technology but in whether someone in the organization owns the decision. A chain with a central purchasing function can relatively quickly introduce a purchasing proposal from AI, because there is already one place that owns that process and has the authority to adjust it. An independent establishment where the owner does the purchasing, planning and service themselves lacks that separation between who assesses the proposal and who does the work — and as a result, AI more often remains an idea rather than a practice there.
Management teams in hospitality formulate ambitions that sound logical on their own: less time on planning, more attention for the guest, faster response to demand. The problem only arises when that ambition is translated one level down. What does "less time on planning" mean for the floor manager who currently puts together the roster themselves? Who will assess the proposal from a system, and does that person have the authority and the time for it? If that translation does not happen, the ambition remains a sentence on a strategy document, while the shop floor is left with a different picture of what is going to change.
That is a pattern that is not unique to hospitality. Other sectors with strong peak moments and a lot of operational work run into similar questions, as can be seen with the AI ambitions in the recreation sector, where seasonal peaks and guest contact play an equally large role. Outside hospitality too, the gap between ambition and shop floor is a recurring theme, for example in how to involve the shop floor in a strategy without the process taking forever. And anyone who notices that a vision statement like "more time for the guest" means something different to everyone may recognize themselves in the question how to make a vague vision concrete.
Ambition without an AI expectation asks different things of an organization than ambition in which AI is already assumed. A chain that expects roster proposals to soon come from a system must first arrange who assesses that proposal, based on what information, and with what authority to deviate. That is not a technical question but a question about decision rights: who is allowed to correct AI's proposal, and is that person equipped for it.
That is usually where it first gets stuck. Not in the technology, but in the question of whether the role that must exercise oversight already exists, or is set up for that task. A floor manager who has never been asked to assess an AI proposal does not automatically know what to look out for or when to say no.
Which work at a specific hospitality business can actually be taken over by AI, which work requires oversight, and which work remains human work, cannot be answered in general terms — it differs by chain, by establishment, by team. That question is answered per task with the work scan from FTE TO AI.
Regarding personnel decisions that might follow from this shift, separate statutory requirements apply for that; that is not a topic on which this page makes statements.
Anyone who wants to know where their own organization currently stands can start with the free readiness check: eight short questions, one per dimension, resulting in a picture of where the organization is furthest along and where the least. The full ambition assessment, with the four layers from vision to target state and the translation to roles and decision rights, is under construction.
Vertel wat u wilt bereiken, dan kijken we samen wat daarvoor moet staan.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.