This week's most interesting audit data came from a property we wouldn't have expected.
Not a one-truck plumber with a 2019 WordPress site. Not a cleaning company with no Google Business Profile.
A 4.7-star Marriott Tribute Portfolio hotel in Washington DC. 161 Google reviews. A lobby bar that shows up in people's trip reports. The kind of property that feels established.
AI visibility score: 25 out of 100.
Here's exactly what that means and why it keeps happening regardless of business size or category.
What the Audit Found
The audit checks five things: Google Business Profile completeness, site health, keyword gaps, missing pages, and AI visibility. On GBP, the hotel scored well — accurate info, strong review volume, recent activity. The problem was entirely in AI infrastructure.
The AI visibility module checks for:
llms.txt — a plain-text file at the website root that briefs AI tools on what your business is, what it offers, and what makes it different. Think of it as a business brief written for AI search rather than human readers. The hotel didn't have one.
Schema markup — structured data that tells search engines and AI tools how to categorize your business. LocalBusiness schema, Hotel schema, Service schema. Without it, AI tools are reading your site without a map. None of it was present.
FAQ schema — the format AI tools use to answer spoken and typed queries. "Is the hotel pet-friendly?" "What's the parking situation?" "How close is it to Georgetown University?" Without FAQ schema, these questions get answered by guessing or pulling from random review text.
AI-friendly meta descriptions — short, structured descriptions that clearly state what you do, where you are, and what makes you different. The hotel's meta description was missing entirely on key pages.
AI knowledge check — we asked Claude directly what it knows about the property. The response: "I don't have specific knowledge of this particular hotel. While I know Tribute Portfolio is Marriott's independent hotel brand, I lack details about this property's offerings and reputation."
That's the critical failure. When a traveler asks an AI assistant for Georgetown hotel recommendations, the AI is working from whatever it has in its training data and whatever it can find in structured form on the web. Without structured data, this property doesn't come up with confidence — even with 4.7 stars and 161 reviews.
Why This Keeps Happening
We've run this audit across tree services, dental practices, HVAC companies, cleaning services, med-spas, law firms, restaurants, and now hotels. The pattern is identical regardless of vertical or business size.
Google ratings don't solve AI visibility. Reviews tell AI tools your business is good. They don't tell it what you offer, where you serve, or what question you answer for the customer.
For a hotel: "good reviews" doesn't tell AI that you're boutique-scale and walkable to Georgetown University. It doesn't tell AI that you have a lobby bar worth mentioning. It doesn't surface you when someone asks about options near the Kennedy Center.
The structured data gap is the same problem we flag for a one-truck plumber: the AI doesn't know how to slot you into its recommendation. A large property on a major travel platform has better distribution than a small local business — but they're still invisible to AI for the same structural reason.
The Fix (Same Three Steps Across Every Vertical)
1. llms.txt at your website root
A plain-text file that describes your property in the same way you'd brief a concierge. For a hotel: property type, location, amenities, room types, dining, what type of traveler you serve best, proximity to landmarks. Written in natural language, not marketing copy.
This is the file AI tools read when they want to understand your business before recommending it.
2. Hotel and LocalBusiness schema markup
JSON-LD structured data on your homepage and key pages that tells search engines and AI exactly what category you're in, where you're located, what you offer, and your contact information. Google's Rich Results Test will tell you if it's valid.
For hotels specifically: add @type: Hotel schema with amenityFeature, starRating, numberOfRooms, checkinTime, checkoutTime. These are the fields AI uses to answer specific traveler queries.
3. FAQ schema on high-traffic questions
The questions travelers actually ask: parking, pet policy, proximity to Georgetown University, C&O Canal access, dining options, check-in time. FAQ schema gives AI tools structured answers to pull from. Without it, the AI has to piece together answers from random review mentions — which it often gets wrong or omits.
The Score Gap Is Consistent
We've now seen businesses with sub-2-star ratings score 60+/100 on AI visibility because they had the technical infrastructure in place. And we've seen 4.7-star properties with packed lobbies score 25/100 because they didn't.
The AI doesn't care about your reputation. It cares about what it can read and categorize.
This is the window. The businesses that build AI infrastructure now are going to have a compounding advantage over the ones that wait until it's obvious why it matters.
If you want to see your current AI visibility score — along with your GBP completeness, site health, keyword gaps, and missing pages — run a free audit at vibetokens.io/start.
Takes 90 seconds. No email required to see the report.
— Murph
