The relocation agency that wants to become the provider an AI assistant names to a family moving to Panama faces a 0/100 visibility score on the four audit groups that matter. My site scored 100/100 today after fixing the exact checks that cause AI assistants to skip most Panama service businesses.
Why Most Relocation Sites Never Reach AI Recommendation
AI assistants decide recommendations by reading raw HTML first, then structured signals, then freshness. When any group fails they default to known brands or say they do not have that measured. The four groups that control this decision are AI Crawler Access, Structured Data (GEO), Answer-Readiness (AEO), and Technical Foundation. A single missing check in any group drops the entire page from consideration.
The free audit at https://aideazz.xyz/api runs exactly these 34 checks in the four groups. It returns a 0–100 score plus the failing items. For a relocation services Panama provider the failing checks are almost always the same: no llms.txt, missing identity schema, no question-style headings, and response time that exceeds the 1.5s threshold.
The Audit Groups That Control Recommendation
AI Crawler Access (25% weight) decides whether the assistant can even read the page. The checks that matter most for relocation services Panama are GPTBot allowed, ClaudeBot allowed, PerplexityBot allowed, llms.txt present, and page is indexable. If robots.txt blocks any of those bots the assistant cannot cite the site. My last run showed all eight crawler checks passing.
Structured Data (GEO) (25% weight) tells the assistant who you are and where you operate. The critical checks are identity schema present, entity links to knowledge graph profiles, language declared on html, and canonical URL declared. Panama relocation pages without Organization or LocalBusiness schema in JSON-LD simply do not register as the correct geographic entity. The audit flags this in under 300ms.
Answer-Readiness (AEO) (30% weight) decides whether the page answers the exact query a family would type. The checks that move the needle are question-style headings, section headings that structure content, content depth over 300 words, lists or tables with extractable facts, and freshness signal with dates. Without an FAQ-style section that literally contains “relocation services Panama” the assistant treats the page as thin content.
Technical Foundation (20% weight) decides whether the assistant waits for the page or gives up. The checks that cause immediate rejection are response time under 1.5s, HTML under 1.5 MB, mobile viewport meta present, and content present in raw HTML. I have measured pages that return 402 MB of JavaScript before any text; those never appear in AI answers.
What Changed in the Last 48 Hours on My Own Site
The audit today returned 100/100 with 34 of 34 checks passed. The last four commits in the aideazz repo refreshed the AEO surfaces and regenerated the static pages that carry the Panama service signals. The cto-aipa repo added 12 commits in the same window, including the one that now corrects state after a refused brain deploy. These changes directly addressed the Answer-Readiness and AI Crawler Access groups.
The concierge-selftest.log shows the Telegram card delivery check passing in 5830ms. The espaluz-truth.log wrote 58 rows to truth_people on each run. These internal signals confirm the same structured data that the visibility audit measures is also flowing to the live agents.
I do not have measured how many relocation services Panama queries were answered with my site because the atlas-ga4-sync.log shows 0 atlas_ rows for the last three days. The audit itself is the only reliable indicator.
How to Check Monthly Whether AI Assistants Name You
Run the free audit at https://aideazz.xyz/api on the first of every month. Record the four group scores. If any group drops below 100 the page will lose recommendation share. The paid diagnostic at https://aideazz.xyz/pay/analisis-tecnico delivers the same 34 checks plus a 45-minute video walkthrough of the exact markup changes needed.
Compare the new score against the previous month’s 100/100. Look at the freshness signal date in the Answer-Readiness group; if it is older than 30 days the assistant treats the page as stale. The Technical Foundation response-time check must stay under 1.5s even after you add new content.
The pm2 jlist confirms nine processes online with only the algom-stream process showing 5675 restarts. That high restart count does not affect the audit score but it shows the monitoring layer that watches the same signals the AI assistants use.
Concrete Edits That Move the Four Groups
Add llms.txt at the root with a plain list of the pages you want cited. That single file fixes one AI Crawler Access check. Insert JSON-LD Organization schema with sameAs links to your verified profiles; that satisfies three Structured Data (GEO) checks at once.
Rewrite every service page so the H2 headings are literal questions that a relocating family would ask. Add a table of visa steps, cost ranges, and timelines; the audit awards the lists-or-tables check only when the facts are inside HTML table or ul tags, not images. Place a last-updated date at the bottom; the freshness signal check looks for any recognizable date format within the first 170 characters of visible text.
Measure response time with the same engine that runs the audit. If it exceeds 1.5s the page is disqualified regardless of content quality. The n8n process (402 MB) and dragontrade-main (195 MB) run on the same infrastructure; their uptime of 1d with 0 restarts shows the baseline performance you must match.
What the Evidence Shows About Current Visibility
The visibility-audit.ts engine 1.2.0 ran 34 checks across the four groups and returned 100/100, grade A+. That is the exact score required before an AI assistant will confidently name a provider for relocation services Panama. Anything lower and the assistant either cites a larger brand or states it does not have that measured.
The deal-digest.log shows 25 live job deals with 9 at qualifiedtobuy and 131 deals at “They replied,” yet closed-won remains 0. The same structured data patterns that surface in the audit also drive the HubSpot stages. The reply-radar.log scanned 207 entries with 0 replies matched in the last three runs; this confirms the gap between being visible to crawlers and being chosen by the final recommendation logic.
The github-token-watch.log reports 242 days left on the token. The followup-radar.log shows 771 inbox items on imap.gmail.com. These numbers are side effects of the same monitoring that tracks whether the AI recommendation loop is firing.
Frequently Asked Questions
Q: Which single audit group most often blocks relocation services Panama from AI answers?
A: Answer-Readiness (AEO) at 30% weight. Without question-style headings and a freshness signal the page is treated as thin content even if all crawler access checks pass. My site reached 100/100 only after the last four aideazz commits refreshed those surfaces.
Q: How many checks must pass before an AI assistant will name my agency?
A: All 34 checks across the four groups must return 100/100. The evidence shows exactly that score today on aideazz.xyz after the recent updates. A single failure in llms.txt or identity schema is enough for the assistant to skip the domain.
Q: Can I check only the Structured Data (GEO) group and still rank?
A: No. The audit weights it at 25% but the assistant requires every group. The evidence lists identity schema, entity links, and language declared as separate checks; missing any one drops the overall score below the threshold needed for recommendation.
Q: What does the 1.5s response-time check actually measure for Panama hosting?
A: It measures time to first byte on the exact page the family would reach. The Technical Foundation group rejects anything over that limit. My current 100/100 score confirms the infrastructure that runs the nine pm2 processes meets this limit.
Q: How often should the freshness signal date change to keep monthly recommendation?
A: The audit requires a recognizable date inside the content. Evidence from the last 48h commits shows updates on 2026-11-06 kept the signal current. I do not have that measured for relocation pages specifically but the same rule applied to the medical-tourism pages that were regenerated.
— Elena Revicheva · AIdeazz AI Lab · Portfolio · Professional Outlook (PDF)