What Does AI Say About Your Property?

AI is describing your hotel to future guests right now.

You didn’t write the description.

You may not have seen it. And it may not be accurate.

There’s an AI-generated summary of your property circulating on the platforms where travelers research before they book. A guest is reading it on their couch while deciding whether you’re worth the click. The description was built from OTA listings, cached blog posts, and reviews you didn’t curate.

Have you checked what it says?

Yesterday on Destination Sunday, I wrote about the version of your town that exists online. The composite picture no one at the destination has assembled. Your property has one too. This is the operator-level version of that same problem.

A TakeUp survey of U.S. leisure travelers found 94% of AI users trust the recommendations at least as much as search engines or travel sites, even as accuracy remains a leading concern.

That combination is the problem. Guests trust what AI tells them. And for independent properties, the AI description is often built from OTA listings, old Booking.com copy, and guest reviews you didn’t curate.

The guest reads someone else’s description of your property and treats it as yours.

When they arrive expecting something you don’t offer, or missing something you do, the stay starts at a deficit. Not because you failed. Because the information that shaped their expectation was never yours to begin with. You won’t notice it in one review. You’ll feel it in your average rating six months from now.

You might be thinking this only matters for larger properties with bigger digital footprints. Or that your particular guests don’t use AI yet. Or that this is one more thing to manage when you’re already stretched across check-ins, maintenance, and the OTA dashboard you opened three tabs ago. Those are fair reactions. But the audit below takes ten minutes, and what it reveals is already shaping your bookings whether you look at it or not.

Five things you can do this week

1. Run the 10-minute property audit

Open an incognito browser window. Google your property name. Read what comes up. Then open an AI tool (ChatGPT, Claude, Perplexity, whatever you have access to) and type: “Tell me about [your property name] in [your town].”

Write down what it says. Note what’s accurate, what’s outdated, and what’s missing entirely. That’s the version of your property that’s shaping booking decisions right now.

2. Compare the AI version to your version

Put the AI description next to your website’s homepage or booking page. Where’s the gap?

Common mismatches: amenities listed that you no longer offer. A description that emphasizes the wrong thing (the AI calls you a “budget-friendly option” when you’ve repositioned as boutique). Room types or pricing language pulled from an OTA listing you haven’t updated in two years. A tone that makes you sound like a chain when you’re a family-run inn.

That gap is where your 4-star reviews come from.

I’ve run this comparison for properties I work with. The mismatch is almost always there. And it almost always traces back to a listing nobody’s touched in two years.

3. Fix your source material

AI reads what’s published. If you want the AI version to change, change what it reads.

Start here:

Google Business Profile. Updates reflect fastest in search and AI results.

Then your OTA listings on Booking.com, Expedia, and Airbnb. Often the primary source AI pulls from for independent properties.

Then your own website. About page, room descriptions, homepage intro.

Start with your Google Business profile. Update the description, hours, photos, and amenities. Then check your OTA listings. If you haven’t touched your Booking.com description since you wrote it three years ago, that’s the version AI is serving to your future guests.

Then update your own website. Your About page, your room descriptions, your homepage intro. Write them in language you’d want an AI to quote back to a potential guest.

4. Check your photo story

Google your property name and click Images. What comes up first?

For many independents, the top photos aren’t from your website. They’re guest uploads from 2021, OTA stock-style shots, or Google Street View from before your renovation. That’s the visual version of the same problem. A guest forms an expectation from photos you didn’t choose.

Upload current, high-quality photos to your Google Business profile. Update your OTA photo galleries. The photos that rank first shape the picture in the guest’s head before your website gets a chance.

5. Ask your front desk what guests expect that you don’t offer

This is the fastest intelligence you have.

Your front-desk staff, your property manager, your check-in team already know where the expectation gap lives. They hear it every week. “I thought you had a pool.” “The listing said you were walking distance to downtown.” “Where’s the restaurant that was recommended?”

That pattern tells you exactly where the online version of your property doesn’t match the real one. Collect it. Write it down. Then trace it back to the source. When you know which listing or description is creating the wrong expectation, you know what to fix. You don’t need a consultant for this. You need a notebook and a Tuesday morning.

Try this prompt

If you want a faster way to spot the gaps, paste this into ChatGPT, Claude, or the AI tool you have access to:

Role
You are a “Public Web Presence Auditor” for hospitality and short-term rentals. Your job is to compare the property’s official description to what a first-time guest will learn from public web sources, and identify trust-breaking mismatches.

Property inputs
- Property type: [type of property]
- Property name: [property name]
- Location: [town/region + country/state]
- Official website URL (when available): [url]
- Official description (paste from homepage or About page):
[PASTE]

Disambiguation rules (must follow)
1) Confirm you are auditing the correct property by matching at least two of:
  - website domain
  - exact address or map pin
  - phone number
  - owner/brand name
  - photos that clearly match the property
2) If you find multiple matches, list them and explain which one you chose and why. Do not proceed until the match is clear.

Sources to check (minimum)
Check and cite the most relevant sources you can find from each category:
A) Google Business Profile / Google Maps listing (priority)
B) Apple Maps or Bing Places when available
C) Major OTAs or marketplaces where this property appears (only those that actually show up in search)
D) Review platforms (Google reviews, TripAdvisor, Yelp, Facebook reviews, etc. as relevant)
E) Social profiles that appear to be official (Instagram, Facebook page, TikTok, LinkedIn when relevant)
F) Local listings (Chamber of Commerce, tourism boards, local directories)
G) Any prominent third-party articles or listicles that rank for the property name

Rules for browsing and accuracy
- Use live web browsing.
- Provide citations with clickable URLs for every factual claim about what the web says.
- Do not guess. If you cannot verify something, label it “Not confirmed.”
- Note “Last checked: [today’s date]” at the top.

Output format (use these headings)
1) Snapshot: What the internet thinks this property is (2–4 sentences)
2) Source log (table):
  - Source | URL | What it claims | Risk (Low/Med/High) | Notes
3) Mismatch report (table):
  - Topic (location, amenities, room types, policies, accessibility, parking, pet policy, check-in, pricing signals, vibe, audience)
  - Official description says
  - Public web says
  - Impact on first-time guest
  - Severity (1–5)
  - Recommended fix
  - Where to fix it (GBP, OTA listing, website, socials, directory)
4) What’s missing that guests expect to know (bullet list, prioritized)
5) Outdated or potentially misleading items (bullet list, prioritized)
6) Quick-win action plan (next 7 days):
  - Top 5 fixes in order, written as tasks
7) Confidence + open questions:
  - What you’re confident about
  - What needs the owner to confirm

‍

Read what comes back. Some of it will be wrong about your property in ways you can trace to a specific listing. That’s your fix list. Some of it will describe your property better than you do. That’s worth noting too. And if your team is skeptical about whether this matters, the output from that prompt is the most concrete conversation starter you’ll find.

You can’t edit the AI. You can edit what it reads.

Expectation misalignment is one of the most expensive invisible problems in independent hospitality. It doesn’t announce itself. It shows up in the gap between what a guest expected and what they found, and it compounds in ratings, in rebooking rates, in the reviews that almost said something great but didn’t.

Most independent operators aren’t checking what AI says about their property. The ones who start will see things they can fix this week. The ones who don’t will keep wondering why bookings feel harder than they should.

When those two versions align, reviews stabilize. When they don’t, you pay for it in your review score.

The guest who doesn’t book never tells you why.

I write about this every week on Substack.

This week’s Destination Sunday: Your Destination Has a Story You’ve Never Read covers the same dynamic at the destination level.

‍
‍

‍

FAQs

How do travelers research destinations before booking?

Most travelers now research destinations through a combination of Google search, AI tools like ChatGPT, Reddit threads, TripAdvisor reviews, and cached blog posts. Marriott Bonvoy’s 2026 research found that 50% of travelers have used AI to plan or research a trip, up from 26% two years prior. Roughly 60% of Google searches end without a click to any external website, meaning many visitors form their impression of a destination entirely from search results and AI-generated summaries without ever visiting the official tourism site.

How accurate is AI-generated travel information?

AI-generated travel recommendations frequently contain errors. A TakeUp survey of U.S. leisure travelers found that while 94% of AI users trust the recommendations, 52% cite inaccuracy as their top complaint. A tourism researcher cited by CNN noted that roughly nine out of ten AI-generated travel itineraries contain mistakes, ranging from recommending closed businesses to fabricating attractions that don’t exist.

How can a destination audit its online information?

A destination information audit takes approximately 20 minutes. Open an incognito browser window and research your town as a first-time visitor would: start with Google, then an AI tool, then Reddit or TripAdvisor. Document what appears, noting outdated listings, closed businesses, wrong hours, and inaccurate descriptions. Fix what you can directly and influence what feeds AI answers by correcting information at the source.

What is a destination information landscape?

A destination’s information landscape is the complete set of online information a potential visitor encounters when researching a place, including search engine results, AI-generated summaries, social media posts, review sites, blog posts, and cached content. Unlike a destination’s brand or marketing, which the destination controls, this landscape is assembled from many sources and is often outdated, incomplete, or written by people with no connection to the community.

‍

‍