The Hotel Operator's Guide to AI

AI for Hotels: How GMs and Owners Actually Use ChatGPT

AI for hotels comes down to one thing: you hand the tool the facts about your property, the context about the guest, and the tone you want, and it hands back a working draft in seconds. In week one, a GM can realistically use AI for guest-review responses, pre-arrival emails, OTA listing descriptions, staff announcements, event inquiries, and shift-report summaries.

1-on-1 with Pete Enestrom, on your actual work

Pete Enestrom

Written by Pete Enestrom

Yale & Columbia, ex-Microsoft & Intel — 1-on-1 AI coaching for executives

Pete Enestrom — executive AI coach and Zaigo co-founder

Your coach’s background

YaleCOLUMBIAUNIVERSITYMicrosoft

AI for hotels is not a lobby robot, a chatbot bolted onto your website, or a revenue-management platform. The tools this guide covers are the general-purpose AI assistants—ChatGPT, Claude, Gemini—and for a hotel operator the honest description is this: an AI assistant is a fast junior copywriter that lives in a chat window. You give it the facts about your property, the context about the guest, and the tone you want, and it gives back a working draft in seconds. Whether you arrived here searching ai for hotels, ai in hospitality, or ai for hospitality, the question underneath is the same—which parts of the writing on my desk can this thing actually take over?

The answer is the writing-and-reading jobs that repeat every week: the guest-review queue, the pre-arrival emails, the OTA listing that hasn't been touched in two years, the staff announcement, the event inquiry sitting in the sales inbox, and the shift notes nobody has time to read carefully. This guide walks through each one with the exact prompts I hand to the hotel operators I coach—written for working GMs and owners, not hotel-tech enthusiasts. Everything here runs on a free plan, and none of it requires changing your property-management system.

What can AI actually do for a hotel in week one?

In week one, an AI assistant gives a hotel GM first drafts of the words the property demands all week long. The response to the three-star review about slow check-in, the pre-arrival email for Friday's arrivals, the reply to the wedding inquiry that came in Tuesday—each of these starts as a blank page, and the assistant removes the blank page. The draft is never the finished product, because the GM still supplies what only the GM has: knowledge of the actual property, the guest's real situation, and judgment about what should and should not be promised. What the assistant hands back is time. The operators I coach describe the first week the same way—the writing that used to eat the evening after the desk quiets down is done before dinner, and the saved hour goes back into the property and the staff.

Week one is also the right time to learn what an AI assistant is not. ChatGPT for a hotel is a writing and thinking partner, not a data source. It has no connection to your PMS, does not know tonight's occupancy or tomorrow's rates, and does not know your brand standards or a guest's stay history. An operator who asks it for local facts—restaurant hours, shuttle times, distances to the convention center—will get confident-sounding answers that may be wrong, and a wrong fact in a guest email becomes tomorrow's bad review. The GMs who get the most from these tools use them the way they would use a bright new assistant: drafts, structure, and wording come from the assistant; facts, figures, and final judgment stay with the operator.

Six hotel prompts worth saving

These are the six prompts I teach first to every hotel operator I coach—starting with the review-response prompt, because the review queue is where most GMs feel the pain. Copy the prompts verbatim, replace the brackets with your own material, and adjust the last line to taste. The prompt is the easy part; your facts and your property's voice are what make the output worth sending.

Guest review responses at scalePrompt

I'm the GM of a [property type—e.g., 80-room boutique hotel / select-service airport property]. Here is a guest review from [Google / TripAdvisor], with the guest's name removed: [paste the review]. Draft a response of 80–110 words: thank the guest, acknowledge the specific issue they raised without making excuses, mention one thing we are doing about it, and invite them back. Match the rating—warm and brief for five stars, apologetic and specific for three stars or below. No guest name, no compensation offers, no promises I haven't approved.

Run your five most recent reviews through this in one sitting. You still approve every response before it posts—the draft is the starting point, not the send button.

Pre-arrival email for the week's arrivalsPrompt

I run the front office at a [property type] in [city]. Write a pre-arrival email for guests checking in this Friday: welcome them, confirm check-in time and parking, mention [seasonal pool hours / the lobby renovation / the restaurant's new menu], and offer help with dinner reservations or a late arrival. Under 150 words, warm and unhurried. Leave [brackets] where my team adds the guest's name and arrival details.

The same skeleton becomes your post-stay thank-you email by changing two lines. Save both versions once and your arrivals and departures are covered for the season.

OTA listing description refreshPrompt

Here are the facts about my property: [room count, room types, amenities, location facts, walking distances I can verify, breakfast/parking/wifi details]. Rewrite our Expedia and Booking.com property description in 200 words. Lead with what guests praise most in our reviews: [e.g., quiet rooms, walkable location, the breakfast]. Describe only what I listed—do not invent amenities, views, or distances. Confident, straightforward tone; no clichés like “hidden gem” or “something for everyone.”

Anything the tool adds beyond your facts is invented until you verify it. Read the draft against your actual amenity list before it goes anywhere near a live listing.

Staff announcement that people actually readPrompt

I'm the GM of a [property type]. Draft a staff announcement about [the holiday scheduling policy / a change to the breakfast shift / a new uniform requirement]. Tone: direct, appreciative, no corporate jargon. Explain what changes, when it starts, why we're doing it in one honest sentence, and who to come to with questions. Under 120 words, suitable for the break-room board and the team group chat.

Read the draft as your longest-tenured housekeeper would. If a sentence would raise eyebrows in the break room, cut it before you post it.

Event inquiry reply from your real factsPrompt

I'm the sales contact at a [property type] with these event spaces: [room names, square footage, capacities, AV, catering options]. A planner wrote asking about hosting [a 120-person wedding next June / a two-day corporate retreat]. Draft a reply: thank them, answer their questions using only my facts, ask the two or three questions I need answered to quote the event, and suggest a site visit. Under 180 words, professional and warm.

Never let the tool quote rates you didn't put in the prompt. Describe your spaces freely; keep pricing in your hands until the inquiry is qualified.

Shift-report summary for the owners' meetingPrompt

Here are my front-desk team's shift notes from the past three days, with guest names removed: [paste the notes]. Summarize them for my Monday owners' meeting in under 200 words: group the issues by theme (maintenance, service, billing), flag anything mentioned more than once, and end with the three items that need a decision this week. Do not soften problems, and do not add issues that are not in the notes.

This one reads instead of writes. It is the fastest way I know to find the pattern hiding in three days of scattered shift notes.

What AI can't do for a hotel operator

An AI assistant cannot see your PMS, cannot check availability, and does not know what happened at your property last night. ChatGPT also writes with total confidence whether it is right or not, which is the dangerous combination in a guest-facing business: a GM who pastes an AI draft into a reply without checking will eventually send a wrong check-out time, an invented amenity, or a shuttle schedule that doesn't exist—and the property, not the tool, owns that review. The working rule I give operators is simple. Facts go into the prompt from you; facts in the output get verified by you. AI for hotels is genuinely useful only when every draft starts from your property facts and ends with your read-through.

Guest reviews deserve their own paragraph, because review platforms and guests both punish canned responses. A reply that could have been posted under any hotel's review reads exactly like that, and a templated apology under a specific, angry complaint does more damage than a late human answer. The assistant's job is to get you from a blank page to an 80-percent draft in thirty seconds; your job is the last 20 percent—the specific detail, the honest sentence, the judgment about whether this reviewer gets a phone call instead. If your property is flagged, add one more step: run your AI use past your management company or brand policy before guest-facing drafts go live. Nothing in this guide is legal advice, and no prompt replaces that review.

What to keep out of the chat window

One short list to print and tape near the front desk. Treat anything you paste into an AI tool as shared with an outside vendor—because that is what it is.

  • Guest names, email addresses, phone numbers, and loyalty account numbers
  • Reservation details, folio charges, or anything in the PMS tied to a person
  • Credit card or payment details of any kind, in any form
  • Incident reports involving identifiable guests or staff
  • Employee personal matters—performance, health, or disciplinary details
  • Anything your brand or management company's written policy hasn't cleared

The mistakes I see hotel operators make with AI

After coaching hotel operators through their first months with these tools, the failure modes are remarkably consistent. None of them are about the technology—they are habits, and every one of them is fixable in a week.

  • One-line prompts. “Write a review response” gets a canned reply any guest can spot. The facts, the rating, and the tone are the brief—and the brief is the work.
  • Posting the first draft. The edit conversation—“shorter,” “less defensive,” “the second sentence over-promises”—is where the quality lives.
  • Pasting guest data. Names, reservation details, and folio numbers do not belong in a chat window. Describe the situation, never the identity; the companion privacy guide on this site goes deeper on where that line sits.
  • Trusting its local facts. The assistant will state a restaurant's hours or a distance to the airport with total confidence and no source. Verify every fact before a guest sees it.
  • Expecting it to know your property. It has never walked your halls. Your notes, your reviews, and your amenity list are the expertise; the tool only shapes the words around them.

Where does 1-on-1 coaching fit?

Everything in this guide, a GM can do alone this week, for free. Self-teaching absolutely works, and most operators should spend a month doing exactly that. The ceiling appears around month two: the generic prompts are running, and the real question becomes which of your workflows—the review queue, the pre-arrival sequence, the way your department heads write to each other—should be rebuilt around these tools, and in what order. That question is specific to your property, your flag, and your team, and generic answers stop helping.

That is the point where working 1-on-1 with a human teacher earns its keep. In a coaching session—we work over video, with operators across the US—we open your actual review queue and your real pre-arrival email, and we build the prompts and habits around your property, your voice, and your brand's rules, not a generic playbook borrowed from the internet. This guide is the map. Coaching is walking it together, on your listings, with your next guest email as the homework.

1-on-1 coaching

See it on your actual workload.

Bring the work you did this week. We'll build the workflows on it together, live.

Format1-on-1Sessions
LocationNYCor Remote
Duration2-4 hrsFocused
ResultClarityWalk away ready
Pete Enestrom in a 1-on-1 coaching session

Your coach

Coached by Pete Enestrom

Yale and Columbia grad, former Microsoft and Intel, and a venture-backed exited founder. Pete has spent the last four and a half years going deep on every major AI tool, and he teaches the way operators learn: on your real work, at your pace, with nothing assumed.

YaleCOLUMBIAUNIVERSITYMicrosoft
Pete Enestrom's signature

Common questions

Straight answers, the way I'd give them across a table.

Keep reading

Beyond coaching

When the workflow needs to be encoded, not just learned — that's what our team at Zaigo builds.