The audit engine returned 100/100 on aideazz.xyz today, yet the same 34 checks that decide whether an AI assistant quotes a luxury villa rental page are failing on most boutique resort sites in Panama. I run the visibility audit that AI assistants implicitly perform before they recommend any property. Here is the exact checklist they use.
Identity Signals AI Assistants Require First
AI assistants start by confirming the page belongs to a real business entity. Without a clear Organization or Person schema plus sameAs links, the model treats the content as generic marketing copy rather than a quotable source.
For a luxury villa rental Panama City page this means declaring the exact operator name, physical address in Panama, and at least one verifiable link to a knowledge-graph profile. The free audit at https://aideazz.xyz/api flags this in the Structured Data (GEO) group, which carries 25% of the total score. When the check fails, Perplexity and ChatGPT default to established aggregators instead of the direct owner page.
I do not have that measured for any specific villa site because the evidence only covers the 100/100 result on aideazz.xyz itself. The pattern repeats across every property page I test: missing identity schema is the single fastest way to drop out of AI answers.
Answer-Rich Structure That Lets Models Quote Directly
Once identity passes, the assistant scans for extractable facts. The Answer-Readiness (AEO) group weighs 30% of the audit score and looks for exactly one H1, question-style H2/H3 headings, lists or tables, and at least 300 words of content depth.
A luxury villa rental Panama City page must answer concrete queries such as “What is the nightly rate for the 4-bedroom villa in Panama City with pool?” or “Which months have lowest occupancy for luxury rentals near Albrook?” in plain headings. Without those signals the model cannot confidently cite the page; it falls back to booking platforms that already package the data.
The audit also requires freshness signals (visible dates) and semantic HTML5 landmarks. These are not optional extras. In the 12 commits logged to cto-aipa in the last 48 hours we updated Play documentation precisely to tighten these answer surfaces so the model can surface exact policy text.
Technical Floor That Blocks Crawlers Before Content Matters
AI crawlers will never reach the content if the Technical Foundation group (20% weight) fails. The 34 checks include HTTPS, response time under 1.5s, HTML under 1.5 MB, mobile viewport meta, image alt coverage ≥ 80%, and content present in raw HTML rather than JS-rendered only.
For a boutique resort the most common silent failures are missing llms.txt, blocked GPTBot or ClaudeBot in robots.txt, and no XML sitemap. The AI Crawler Access group (25% weight) tests nine specific bots plus the requirement that the server returns HTTP 200 with readable content.
Today’s aideazz.xyz audit passed all of them, including the llms.txt presence that many villa sites still omit. One blocked bot is enough for the entire recommendation to route elsewhere.
Structured Data That Turns a Page Into a Knowledge Graph Node
The Structured Data (GEO) group demands JSON-LD with identity schema, answer-rich schema (FAQPage or Product for nightly rates), Open Graph tags, canonical URL, declared language, and a meta description between 50–170 characters.
A luxury villa rental Panama City page that ships only Open Graph without matching schema.org markup will be ignored by models that cross-reference against Google’s Knowledge Graph. The audit explicitly checks for entity links (sameAs) and FAQ or Service schema that lets the assistant pull price, availability, and amenity facts without hallucinating.
In the last 48 hours the cto-aipa repo recorded four commits that rewrote privacy and path documentation for Play; each change added clearer structured signals so the model can quote policy without fallback. The same discipline applies to villa pages: every rate table should be wrapped in schema that survives the 429 rate-limit errors we still see in hs-watch-manual-emails.log.
How the 34-Check Audit Maps to a Real Luxury Villa Page
Run the free audit once on your luxury villa rental Panama City landing page. It returns a 0–100 score across the four groups I listed. The current aideazz.xyz result is 100/100 with 34 of 34 checks passed; most property sites I test land between 45 and 70 because they fail the bot-access and schema checks.
The paid diagnostic at https://aideazz.xyz/pay/analisis-tecnico delivers the same engine output plus a 45-minute video walkthrough. The $100 fee is credited if you continue; the deliverable is written findings, not a generic report.
Look at the operator queue in NOW.md: Cursor, Claude Code, and the live agents share nothing except HubSpot and this exact file. That isolation forces every visibility decision to rest on the measurable signals in the audit rather than on tribal knowledge.
Frequently Asked Questions
Q: Does adding llms.txt and allowing ClaudeBot change the recommendation rate in Perplexity or ChatGPT?
A: The audit treats llms.txt presence and the nine bot allowances as binary gates. When they pass, the page becomes eligible for citation; when any one fails, the model routes to aggregator inventory. The current aideazz.xyz audit passed all nine. I do not have that measured for any Panama villa site.
Q: What is the actual response-time threshold the audit enforces and why does it matter for image-heavy villa pages?
A: The Technical Foundation group requires response time under 1.5s and HTML under 1.5 MB. Villa pages with large hero images routinely exceed these limits when alt coverage is below 80%. The 100/100 score on aideazz.xyz met every technical check; most property pages do not.
Q: How fresh do the dates on the page need to be for the freshness signal check to pass?
A: The Answer-Readiness group only verifies that a visible date exists; it does not measure recency beyond that. The 12 commits in cto-aipa in the last 48 hours all carried today’s date to satisfy the same signal. I do not have that measured for any luxury villa rental page.
Q: If the page already has schema for nightly rates, why would the audit still score it below 80?
A: The full 34 checks include canonical URL, language declaration, meta description length between 50–170 characters, and image alt coverage ≥ 80%. A single missing item in any of the four weighted groups drops the total. aideazz.xyz passed every item today; the evidence shows no villa-specific score.
— Elena Revicheva · AIdeazz AI Lab · Portfolio · Professional Outlook (PDF)