Research

5,913 websites. Grade F. The Central Europe AI Readiness Study 2026.

We analysed the structured data of Europe's largest websites — so you don't have to. This page gives you an early look at what we found: who's doing it right, who's leaving visibility on the table, and what the data says about the state of schema markup across the continent.

// Study parameters

The dataset.

Scope	5,913 domains · 9,275 URLs analyzed · Snapshot: 6 March 2026
Method	3-C Framework — Content (40%) · Code (35%) · Credibility (25%)
Tool	Puppeteer Headless Browser — Schema extraction, content analysis, technical audit, Core Web Vitals
Validation	SUTSCHE/Boye Study — Independent validation on 1,000+ pages from DAX, SDAX, FTSE 250, SMIM/SPI, AMX

// The paradox

The infrastructure is there. The signal isn't.

Fast servers. Zero content signals. No credibility markers. Central European websites score excellent on Core Web Vitals — and near-zero on every dimension AI systems use to evaluate, understand, and cite content.

// Core Web Vitals

Technically excellent.

648ms

LCP

Largest Contentful Paint — good

500ms

FCP

First Contentful Paint — good

0

CLS

Cumulative Layout Shift — perfect

97ms

TTFB

Time to First Byte — excellent

// Key findings

The numbers don't lie.

27

Median 3-C Score

Out of 100. Grade F. Infrastructure without signal.

0%

Credibility signals

Author attribution. Publication dates. Cited sources. On zero URLs.

3.5%

Schema-rich domains

Only 3.5% have schema rich enough to be useful to AI systems.

49%

AI bots blocked

URLs that partially or fully block AI crawlers.

// 3-C Framework

Three dimensions. One score.

Score = (Content × 0.40) + (Code × 0.35) + (Credibility × 0.25)

25

Content — 25/100

What is being said? Statistics, quotes, source references, comparison tables — absent across the board.

45

Code — 45/100

How is it structured? Schema.org, Technical SEO. Best dimension — still below passing.

5

Credibility — 5/100

Who says it? Authors, dates, external references. Near-zero on every metric.

// Deep dive: Schema saturation

66.9% have no schema. Of those that do, most is empty.

// Schema quality

Six tiers of schema quality.

None	3,954 domains (66.9%) — No schema markup at all
Broken	1,143 domains (19.3%) — Types declared, properties empty (0% completeness)
Skeleton	71 domains (1.2%) — Less than 40% of properties filled
Thin	102 domains (1.7%) — 40–60% property completeness
Medium	434 domains (7.3%) — 60–80% property completeness
Rich	209 domains (3.5%) — 80%+ completeness, 3+ types — actually useful for AI

Why 19.3% are "broken": 1,143 domains have Schema.org types in their HTML, but no parseable properties. Main causes: 405 domains with Article schema — CMS-generated, properties empty. 169 domains with Organization without usable properties. These sites think they have schema. They don't.

// Critical gaps

The five hardest property findings.

99%	mainEntity on WebPage — missing. The semantic link between page and content. Without it, AI can't connect your page to what it describes.
78%	description on Organization — missing. Who is this company? AI systems can't answer.
63%	publisher on WebSite — missing. Who owns this site? Unknown to machines.
51%	sameAs on Organization — missing. No external identity anchoring (Wikipedia, LinkedIn, social profiles).
97%	jobTitle on Person — missing. Author authority completely invisible.

// Type combinations

How types are combined.

Ideal	89 domains (1.5%) — Organization + WebSite + WebPage + Breadcrumb
Good	266 domains (4.5%) — At least Organization + WebSite
Partial	487 domains (8.2%) — Only Article (without organizational context)
Minimal	166 domains (2.8%) — Only Organization alone

// Deep dive: Credibility

Trust signals. All zero.

Author	0% — Author attribution on analyzed URLs
Date	0% — Publication date present
Current	0% — Current year marker
Sources	0% — Cited sources or references
Org Schema	10% — Organization schema present (but mostly empty)

// Deep dive: Content

Content signals. Nearly absent.

Statistics	0% — Pages with statistical data
Quotes	0% — Pages with cited quotes
Sources	0% — Pages with source references
Tables	0% — Pages with comparison tables
Pro/Con	0% — Pages with structured arguments
Word count	13% — Pages exceeding 1,000 words. Length without substance.

// AI bot access

Who's allowed in. Who's not.

GPTBot, ClaudeBot, Google-Extended, Bingbot, PerplexityBot analyzed across 5,913 domains. Not a deliberate strategy. Nobody checked.

51%

Fully allowed

Open access for all analyzed AI bots.

43%

Partially blocked

Some AI bots blocked, others allowed. Inconsistent configuration.

6%

Fully blocked

All analyzed AI bots blocked via robots.txt.

// Independent validation

An independent study reached the same conclusions.

SUTSCHE GmbH — an independent DXP consultancy with no vendor contracts — analyzed 1,000+ pages from DAX, SDAX, FTSE 250, SMIM/SPI, and AMX companies using the same 3-C Framework. Their findings mirror ours.

Countries: UK (283), Germany (276), Switzerland (159), Netherlands (153), Denmark (125), Belgium (101), Austria (98), Luxembourg (58).

// SUTSCHE/Boye findings

Key results.

No Schema	47% — No schema markup at all
FAQPage	0.9% — FAQPage adoption
Article	3% — Article schema present
Author	1% — Author schema present
AI blocked	27% — AI bots actively blocked
AI open	38% — Fully open for AI bots
No date	80% — No publication date
No H1	18% — Missing H1 heading

// Methodology

How we measured.

enhancely Study: 5,913 domains, 9,275 URLs, snapshot 6 March 2026. Puppeteer Headless Browser. Schema extraction + property analysis (5,159 raw files). Content analysis: GEO signals (7 categories). Technical SEO: 10 signal types. Core Web Vitals: synthetic measurement, 2,263 URLs. Scoring: 3-C Framework.

Scientific foundation: Assessment framework based on synthesis of 16 peer-reviewed papers and industry reports (Aggarwal et al. KDD '24 · Dang et al. Semantic Web 2025 · Chen et al. 2025 and others).

Limitation: Synthetic measurements. Breadth study (1 URL/domain). No deep crawl. Core Web Vitals: Headless Chrome, no real user data.

// Literature

Research sources.

KDD

GEO: Generative Engine Optimization

Aggarwal et al., 2024

SWJ

LLM4Schema.org

Dang et al., Semantic Web Journal, 2025

AI

Search Engines in an AI Era

Venkit et al., 2024

Vis

Manipulating LLMs for Product Visibility

Kumar & Lakkaraju, Harvard, 2024

// FAQ

Questions about the study.

The 3-C Framework evaluates AI readiness across three dimensions: Content (40% weight) — measures GEO signals like statistics, cited sources, comparison tables, and structured arguments. Code (35% weight) — measures Schema.org coverage, technical SEO, and structured data quality. Credibility (25% weight) — measures author attribution, publication dates, external references, and organizational identity signals. The composite score is calculated as (Content × 0.40) + (Code × 0.35) + (Credibility × 0.25), resulting in a 0–100 scale. The median score across 5,913 Central European domains was 27/100 — Grade F.

Most CMS platforms do not generate Schema.org markup by default. WordPress, TYPO3, and other enterprise systems produce HTML that is visually correct but structurally empty for machines. SEO plugins like Yoast cover only 20–30 generic types and require manual configuration per page. The result: two-thirds of European websites have zero structured data, and another 19.3% have broken schema — types declared but properties left empty by CMS auto-generation. Only 3.5% of analyzed domains have schema rich enough (80%+ property completeness, 3+ types) to be useful for AI systems.

SUTSCHE GmbH — an independent DXP consultancy with no vendor relationship to enhancely — analyzed 1,000+ pages from major stock index companies (DAX, SDAX, FTSE 250, SMIM/SPI, AMX) across 8 countries using the same 3-C Framework. Their findings mirror the enhancely study: 47% have no schema at all, 0.9% FAQPage adoption, 1% Author schema, 80% missing publication dates, and 27% actively blocking AI bots. Two independent datasets, one conclusion.

If your website is hosted in Central Europe, there is a 96.5% probability that your Schema.org markup is insufficient for AI systems — either missing entirely (66.9%), broken (19.3%), or too thin to be useful (10.2%). The good news: because so few competitors have adequate structured data, implementing comprehensive Schema.org markup now creates a significant competitive advantage. enhancely can analyze your domain in under 60 seconds and show exactly where you stand on the 3-C scale.

Credibility signals in the 3-C Framework refer to machine-readable markers — not human-visible content. A website may display an author name on a blog post, but without Person schema with jobTitle, sameAs (linking to LinkedIn/Wikipedia), and affiliation properties, AI systems cannot verify or weight that authority. The study found 0% author attribution, 0% publication dates, and 0% cited sources in structured data across all analyzed URLs. The information exists on pages — it is just invisible to machines because it lacks Schema.org markup.

Your content is already perfect for humans.

Get started with enhancely