AI in hospitality: use cases across the guest journey

Most coverage of AI in hospitality reads like a feature list. Chatbots, dynamic pricing, smart rooms, voice agents, personalization engines, all stacked up as if they were the same kind of thing. They are not. They show up at different points in the guest journey, they solve different problems, and they fail in different ways.

A more useful way to think about it: walk the journey a guest actually takes, from the first search to the post-checkout email, and ask at each stage what AI is genuinely good at and where it gets in the way. That is the frame this piece uses. The goal is to help you spend attention on the places AI pays off and skip the places it just adds noise.

The guest journey, in four stages

A guest moves through roughly four stages: discovery and booking, pre-arrival, the on-property stay, and post-stay. AI touches all four, but its job is different in each.

In the early stages, AI is mostly about answering and converting. In the middle, it is about service and logistics. At the end, it is about follow-up and learning. The mistake is assuming one tool does all of it well. A booking assistant and an in-room concierge are solving genuinely different problems, even if a vendor sells them under the same banner. Generative AI in particular helps unevenly across the journey; we map where it improves the guest experience on its own.

The throughline is that guests increasingly want to self-serve the routine parts and reach a human for the rest. In Oracle and Skift's Hospitality in 2025 study, which surveyed 5,266 consumers and 633 hotel executives in spring 2022, 77% of travelers said they were interested in using automated messaging or chatbots for customer service requests, and 73% said they were more likely to stay somewhere offering self-service technology. That preference is the engine behind every stage below.

Stage one: discovery and booking

This is where AI earns the most direct revenue and where the category is shifting fastest. A guest researching a stay asks the kind of open question a static booking form cannot handle: is this hotel good for a family with a toddler, is the gym actually decent, can I get a room away from the elevator. A conversational assistant grounded in real property information can answer those and move the guest toward a direct booking instead of bouncing them to an OTA.

The pull toward direct booking is the commercial point. Every reservation captured on the hotel's own channel avoids the commission an online travel agency takes, so even a modest lift in direct-booking conversion shows up clearly in the margin. That is why booking assistants tend to be the first AI investment a revenue-minded property makes, and the same conversational layer handles the changes and cancellations that come after the booking exists.

The risk at this stage is accuracy. An assistant that invents an amenity or quotes a rate that does not exist creates a problem that surfaces at check-in, when the guest is standing at the desk holding you to a promise the AI made. Grounding the assistant in live, correct property and rate data matters more than how fluent it sounds.

Stage two: pre-arrival

Between booking and arrival, the work is logistical. Guests want to confirm details, ask about parking and check-in, add a request, maybe accept an upsell to a better room or an early check-in. This is high-volume, low-complexity messaging, and it is exactly what AI handles well.

It is also where proactive outreach pays off. A pre-arrival message that confirms the booking, offers a room upgrade, and answers the predictable questions in advance does two things at once. It captures incremental revenue, and it removes those same questions from the front desk before they ever arrive. The personalization gap here is real: Medallia's 2024 study of 1,749 hotel guests found that 61% of consumers would spend more for a personalized experience while only 23% said their recent hotel stays felt highly personalized. Guests will pay for personalization, and delivery has been lagging that demand.

Channels matter at this stage because guests answer where they already are. WhatsApp and SMS see far higher engagement than email for pre-arrival messages, which is why hospitality messaging keeps gravitating to those channels. Getting WhatsApp customer support for travel and hospitality right comes down to the channel mechanics.

Stage three: the on-property stay

This is the stage everyone pictures, and it is the one where AI is most oversold. The fantasy is a fully automated guest experience. The reality is narrower and more useful.

On property, AI is good at absorbing the steady stream of small requests that otherwise interrupt the front desk: what time does the pool close, can I get extra towels, book me a table for two at eight, what is the WiFi password. HiJiffy, a hospitality messaging vendor, reports its AI autonomously resolves over 85% of incoming guest queries across more than 2,100 hotels, and the dominant themes are reservations, amenities, policies, and general information. Those are the answerable requests, and offloading them frees staff for the guest standing at the desk with a real problem. The mechanics of pointing AI at that pile of interruptions are their own guide: using AI to cut front-desk load at hotels.

Where AI should stay out of the way on property is anything emotional or unusual. A guest whose room is not ready, who has a complaint about noise, who needs a comp, who is having a bad trip: those moments are where a hotel earns loyalty or loses it, and they need a person. A system that recognizes its own limits and hands off cleanly when it is unsure, the way our Agent 5 model does, beats one that tries to talk its way through a situation it cannot read. Guests forgive an AI that fetches a human. They do not forgive one that mishandles a bad moment.

The honest version of in-stay AI is a quiet utility that handles the routine and gets out of the way for the rest. The 2025 State of Hotel Guest Tech Report, a survey of 402 recent hotel guests, found 70% of guests find chatbots helpful for simple requests like the WiFi password and room service, which is precisely the scope where it works.

Stage four: post-stay

After checkout, AI does two jobs. It runs the follow-up (feedback requests, review prompts, loyalty nudges, win-back offers), and it learns from the stay. Conversational feedback that asks a guest a real follow-up question in their own words surfaces the actual driver of a complaint or a compliment, which a five-star scale never captures.

This is the least glamorous stage and one of the most valuable, because it feeds the rest. The post-stay data is what tells you which pre-arrival upsells convert, which in-stay questions spike, and where the AI is getting things wrong. AI in hospitality works as a loop, where what you learn after checkout improves what happens before the next booking.

What this means for how you buy

Looking at the journey as four stages changes the buying question. Instead of asking "should we get hospitality AI," you ask "which stage is costing us the most right now, and what is the narrowest thing that fixes it."

A property bleeding commissions starts at booking. A property with a buried front desk starts in-stay. A property with flat loyalty starts post-stay. Adjacent travel verticals feel the same pressure in sharper form, like airlines handling disruptions, rebooking, and baggage when a storm multiplies contact volume in an afternoon. The stages are connected, so a tool that covers the guest's messaging across all of them keeps the context consistent as the guest moves through. The investment case is always clearest when it is anchored to one stage's specific pain. A vague promise to transform the experience everywhere at once is much harder to hold a vendor to.

The hotels that get the most out of AI tend to share a habit: they are precise about which moments belong to the machine and which belong to a person. The pool-hours question belongs to the machine. The guest in tears at midnight belongs to a person. Get that division right at each stage of the journey, and AI stops being a gadget and starts being the thing that gives your staff room to do the part only they can.