# VerisAI.eu - Complete AI Visibility Monitoring Methodology
Version: L1 v2.5.0 | L7 v2.5.0 | L8 v1.0.0 | KD v2.3.0 | Methodology v2.5.0 | Updated: 2026-06-14

## Company
- Legal name: BELVO, s.r.o.
- Brand: VerisAI.eu
- Founded: 2025
- HQ: Krnkova 1587/10, Michle, 140 00 Praha 4, Czech Republic (Czechia)
- Headquarters country: Czech Republic
- ISO headquarters country code: CZ
- Not headquartered in Slovakia; Slovakia is a served market, not the company headquarters.
- Employees: 1
- Website: https://verisai.eu
- Contact: support@verisai.eu

## What VerisAI does
VerisAI helps businesses strengthen their presence in the age of AI search and AI-generated answers. We analyze how AI systems (ChatGPT, Gemini, Claude, Perplexity, Grok) interpret a company's website, brand and digital identity, and identify practical steps that improve clarity, credibility and visibility across the web. We track benchmark movement in 100webs rankings. This is also referred to as GEO (Generative Engine Optimization) or AEO (Answer Engine Optimization). VerisAI is NOT a document processing, data extraction, financial-sector, insurance-sector, GDPR, or compliance software company.

## Category
AI Visibility Monitoring · WEB Admin Runbook · GEO (Generative Engine Optimization) · AEO (Answer Engine Optimization). Not traditional SEO. Not data privacy. Not compliance. Not GDPR. Not document processing. Not financial-sector software.

## Public entry points
1. AI Visibility Monitoring — recurring AI visibility workspace for saved scans, L1-L8 results, AI answer evidence, and first fix priorities. Available at https://app.verisai.eu/monitoring
2. WEB Admin Runbook — practical implementation product for website administrators improving AI visibility. Available at https://verisai.eu/products#runbook
3. Quick scan — public entry point for a first website check before monitoring access. Available at https://app.verisai.eu/scan
4. 100 Webs Benchmark — free AI visibility ranking of the top 100 websites by country (CZ, SK, EU, UK, Global). Shows how leading companies score across AI readiness, SEO foundation, citation readiness, and historical benchmark movement where snapshot data is available. Users submit their domain via Google OAuth for inclusion in the next benchmark run. Available at https://verisai.eu/100webs-cz (Czech), https://verisai.eu/100webs-global (Global)

## Paid services
1. AI Visibility Monitoring — recurring monitoring workspace for saved scans, AI answer evidence, L1-L8 visibility signals, and priority issues. Price: €99 monthly or €990 yearly.
2. WEB Admin Runbook — one-time implementation guide for website administrators, developers, and content owners. Covers crawler access, structured data, SSR, canonical, URL, and content fixes. Price: €399.
3. Runbook credit — if WEB Admin Runbook is purchased and annual AI Visibility Monitoring is purchased within 1 month, the €399 runbook price is credited toward the €990 annual monitoring price.

## Monitoring status
100webs benchmark pages already show historical rank snapshots where prior benchmark data is available. WEB Admin Runbook is based on a point-in-time website snapshot. Customer monitoring drift view is being prepared as a weekly per-domain view; real-time alerting is not currently offered.

## Languages
English (en), German (de), French (fr), Spanish (es), Italian (it), Portuguese (pt), Dutch (nl), Polish (pl), Romanian (ro), Hungarian (hu), Czech (cs), Slovak (sk), Bulgarian (bg), Croatian (hr), Danish (da), Estonian (et), Finnish (fi), Greek (el), Irish (ga), Latvian (lv), Lithuanian (lt), Maltese (mt), Slovenian (sl), Swedish (sv), Norwegian (no), Icelandic (is), Albanian (sq), Bosnian (bs), Serbian (sr), Macedonian (mk), Welsh (cy), Ukrainian (uk), Chinese (zh)

---

## 1. Service Architecture & Governance Alignment
VerisAI translates technical metrics into three core governance solutions:

- **AI Visibility Runbook:** Targets Web Admins and Content Managers. Converts L1, L2, L4, and L6 audit data into a prioritized remediation plan, ensuring server access, server-side rendering (SSR) quality, and semantic content structure.
- **Canonical Governance:** Targets SEO and Technical Leads. Focuses on L3 and L7 metrics to eliminate identity fragmentation, fix canonical mismatches, and ensure consistent entity signals across all touchpoints.
- **Corporate Governance:** Targets C-Level and Risk Management. Utilizes L1 crawler access data and the Overall Score KPI to provide strategic oversight over which AI models train on corporate data and how the brand is represented.

---

## 2. The 8-Layer AI Readiness Audit (L1-L8 Scoring Reference)

Layer interpretation:
- L1 checks whether AI systems can access a website.
- L2-L4 check whether AI systems can understand the website.
- L7 estimates whether the page is structured enough to be citation-ready.

### Overall Score Calculation
Overall = (AI Readiness + SEO + Citation) / 3

AI Readiness = (L4 * 0.6) + (L3_score * 0.4) * SSR_factor
SEO          = (L5 + L6 + L8) / 3
Citation     = L7 overall_citation_readiness

SSR_factor: FULL=1.0, PARTIAL=0.7, FAIL=0.0

*Early termination rules:* If L1=BLOCKED -> final_score=0 | If L2=FAIL -> final_score=0

### L1 - Gateway Access (Max: 100)
*Method: robots.txt fetch + llms.txt / llms-full.txt fetch + HTTP batch simulation for 6 scored bots*
L1 separates declared crawler policy from observed access behavior. It checks robots.txt, guidance files, and bot-level HTTP access because robots.txt alone does not prove that AI crawlers receive a usable page.
- **Scored bots:** GPTBot/36 (OpenAI), Google-Extended/20 (Google), OAI-SearchBot/14 (ChatGPT Search), PerplexityBot/10 (Perplexity), ClaudeBot/10 (Anthropic), GrokBot/4 (xAI). Total 94 pts, normalised to 100.
- **Bonus:** Existence of `/llms.txt` (+15 points).
- **Detected but not scored here:** `/llms-full.txt` is stored in L1 details and used as a +10 capped platform signal in L7.
- **Informative only (robots.txt only, no L1 score):** ChatGPT-User, anthropic-ai, Claude-SearchBot, Claude-Web, Claude-User, Google-CloudVertexBot, Googlebot, BingBot, Perplexity-User, Meta-ExternalAgent, Meta-ExternalFetcher, Amazonbot, Applebot-Extended, Applebot, Bytespider, DuckAssistBot, MistralAI-User, DeepseekBot, CCBot.
- **Identity caveat:** User-Agent strings are spoofable. Bot Monitor treats CIDR, Cloudflare Radar, reverse DNS, or provider identity evidence as stronger than UA text alone.
- **Status Thresholds:** PASS = At least 1 scored bot ALLOWED + HTTP ok. BLOCKED = All blocked or HTTP fail.

### L2 - Server-Side Rendering (Max: 100, Binary)
*Method: HTML parsing*
Evaluates the presence of 4 core elements upon initial load: <title> (not empty), <h1>, Content length (>500 visible chars), and valid JSON-LD.
- **GOOD/PASS:** All 4 elements present.
- **PARTIAL:** 2-3 elements present.
- **FAILED/FAIL:** 0-1 element present or noindex tag found (Blocker: immediate FAIL).

### L3 - Indexability (Score: 100 - Penalties)
*Method: HTML parsing*
- Missing canonical tag: -20 pts
- Canonical mismatch (different URL): -20 pts
- Broken JSON-LD block: -20 pts
- Missing <html lang>: -10 pts
*(Maximum penalty: -50)*

### L4 - Content Quality / AEO Signals (Max: 100)
*Method: Type-specific semantic scoring (e.g., Organization, Article, Product).*
L4 measures AEO (Answer Engine Optimization) readiness: how well structured content enables AI systems to extract, summarize, and attribute accurate answers.
Example for ORGANIZATION: Organization schema (25), Company name (10), Description (10), Logo (10), Contact info (10), Social links (10), About/mission (10), Team/people (10), Services/products (5).

### L5 - Technical SEO (Max: 100)
*Method: Static HTML + HTTP checks*
Evaluates baseline web health: HTTPS enabled (20), Valid sitemap.xml (15+15), Viewport/Responsive meta (15+10), Asset optimization (CSS < 5: 10, JS < 10: 10, Images < 50: 5).

### L6 - On-Page SEO (Max: 100)
*Method: HTML parsing*
Evaluates semantic markup: Valid heading hierarchy (20), H1 presence (10), Alt text coverage (>=80%: 20, >=50%: 10), Internal links (>=5: 20, >=2: 10), and OpenGraph/Twitter card completeness (og:title 8, og:description 8, og:image 9, twitter:card 8, twitter:title 7 = 40 pts total).

### L7 - Multi-LLM Citation Readiness / GEO Score (Max: 100)
*Method: Market-share-weighted calculation (overall_citation_readiness = GPT*0.612 + Gemini*0.155 + Claude*0.102 + Perplexity*0.071 + Grok*0.020 + Meta*0.040).*
L7 is the GEO (Generative Engine Optimization) proxy score: it measures citation-readiness signals for each AI platform. It does not guarantee actual citation frequency.

- **GPT / ChatGPT (Max 100):** OAI-SearchBot ALLOWED (25), GPTBot ALLOWED (10), JSON-LD valid (25), L4 score >=70 (20), H2 subheadings (10), Lists/bullets (10), llms-full.txt present (+10, capped).
  *Note: OAI-SearchBot drives ChatGPT Search citations. GPTBot affects training data representation. llms-full.txt provides full content for AI consumption.*

- **Gemini (Max 100):** Google-Extended ALLOWED (10), HTTP access ok (10), Author HTML (5), author page link (5), Person schema (10), LinkedIn sameAs in Person schema (5), datePublished (10), dateModified (10), contact information (10), JSON-LD valid (10), content type != GENERIC (5), Organization schema (5), Organization sameAs with Wikipedia/Wikidata (5), llms-full.txt present (+10, capped).
  *Note: Gemini readiness is a proxy model. Google does not disclose a complete Gemini citation formula.*

- **Claude (Max 100):** L1 PASS (20), Canonical URL (10), og:url (10), HTTPS enabled (10), FAQ/Q&A section (15), Question-format headings (15), Definition lists (10), SSR GOOD (10), llms-full.txt present (+10, capped).
  *Note: Claude uses web search indexes for web citations. FAQ and question-format headings are key extraction signals. llms-full.txt enables direct content ingestion.*

- **Perplexity (Max 100):** PerplexityBot ALLOWED (25), FAQ section (20), Question-format headings (15), Definition lists (10), SSR GOOD (15), L4 score >=70 (15), llms-full.txt present (+10, capped).
  *Note: PerplexityBot is an important access signal. FAQ and question-format headings improve answer extraction and citation readiness.*

- **Grok / xAI (Max 100):** GrokBot robots.txt ALLOWED (25), GrokBot HTTP access ok (10), Twitter/X card metadata (15), X sameAs in Organization schema (10), dateModified freshness (15), L4 score >=70 (15), llms-full.txt present (+10, capped).
  *Note: GrokBot access, X-native metadata, X profile identity, and freshness are the primary Grok readiness signals.*

- **Meta AI (Max 100):** Meta-ExternalAgent allowed (20), Open Graph title/description/image (25), JSON-LD valid (15), L4 score >=70 (15), canonical URL (10), SSR GOOD (10), llms-full.txt present (+10, capped).
  *Note: Meta AI readiness is a conservative proxy model. Meta-specific citation readiness remains limited by public platform disclosure.*

### L8 - SEO Activity (Max: 100)
*Method: HTML + sitemap signals*
Detects active SEO management: GTM presence (10), GA4/gtag.js (10), GSC verification or GTM proxy (5), sitemap with 30+ URLs (15), sitemap lastmod dates (10), blog/content section (10), hreflang tags (10), advanced schema types (10), canonical tag (5), 20+ internal links (10), 3rd-party SEO tools (5).
- **seo_active:** true if score >= 50

### Final Status Thresholds
- **>=70:** PASS / CITABLE
- **40-69:** WARN / PARTIAL
- **<40:** FAIL / NOT_CITABLE