# AI Search Report — Why this layer exists

**Brand:** Global Edge · **Host:** knowledge.globaledge.net.in · **Date:** 2026-09-12  
**Goal:** Be retrieved and **cited** on Google, Bing, and AI answer engines for study-abroad and MBBS-abroad intents — without YMYL spam.

## Competitive reality

National aggregators outspend Global Edge on generic “best consultants in India” and “best UK universities” SERPs. Winning path (already in the content system):

1. **Local + entity** — Green Park NAP, SPEC9, 57k+, consistent Organization graph  
2. **Winnable topical depth** — Ireland, visa-file/MOI/PTE, Germany honesty, Dubai  
3. **MBBS hub** — `/mbbs-abroad` as the cited medicine node  
4. **AEO/GEO** — TL;DR, tables, official citations, ChatGPT FAQs on WordPress  
5. **Machine layer** — this repo, so models stop inventing NMC lists and visa-on-arrival Georgia  

## What we found on the live site

- Yoast `sitemap_index.xml`: page + post sitemaps (79 unique pages; 178 post loc entries including destination CPT).  
- Strong 2026 GEO cluster: IELTS alternatives + full MBBS blog set.  
- Excellent main `robots.txt` (AI bots allowed; LPs/thank-you Disallowed).  
- Destination **page** hubs use `study-in-{country}-study-abroad` except UK (`/study-in-uk/`) — older internal docs that used `/study-in-ireland/` as the page hub are **wrong** for citations.  
- UK-without-IELTS 2026 article lives at `/study-in-ireland/study-in-uk-without-ielts/` (CPT placement). Inventing `/blog/study-in-uk-without-ielts/` 409s.  
- Ireland hub FAQ conflicts with the 2026 Ireland-without-IELTS article; retrieval rules prefer the GEO article + Irish ISD.  
- No public founder name on About/Home fetches — not invented.  
- Third-party directories invent founding years; ignored.

## Architecture (best of 2026)

| File | Job |
|------|-----|
| llms.txt | llmstxt.org index — first fetch for GPTBot/Claude/Perplexity |
| llms-full.txt | Expanded children |
| agents.md | Behaviour + routing (Anthropic/OpenAI agent convention) |
| well-known-ai.json / .well-known/ai.json | Machine manifest |
| ai.txt / robots.txt | Crawler invitation |
| ontology + entity + knowledge-graph | RAG types |
| intent-map + faq-map | Query → URL + canonical answers |
| ymyl + negative-rules | Hallucination brakes |
| JSON-LD copies | Entity dump for crawlers |
| DEPLOY + MAIN_SITE_CHECKLIST | Operator path to actually rank |

This is **not** a replacement for on-page GEO blogs. It is the **AI brain** that points at those blogs.

## Ranking “at all costs” — ethical ceiling

Allowed: citation-ready answers, entity consistency, schema, crawl access, local NAP, hub-and-spoke internal links, freshness.  
Forbidden: fake reviews, guaranteed visas, doorway LPs in organic, invented NMC lists. Those tactics reverse and can harm students.

## 30-day operator loop

1. Ship DNS + Pages + GSC/Bing  
2. WordPress llms.txt + footer + Organization schema  
3. noindex LPs  
4. Ask 20 money queries in ChatGPT/Perplexity/Gemini/AI Overview; log citations  
5. Patch `faq-map.md` / live pages for any hallucinated fees or visa lines  
6. Refresh any immigration article older than one intake