combined GPT + Gemini weight in the current VerisAI citation-readiness model.
EU AI Readiness Observatory · Data Note 02
83% of the AI market.
Only 17.4% ready for both.
GPT and Gemini represent 83.0% of VerisAI’s current market-share weighting. Yet only 17.4% of 63,858 crawler-accessible EU-27 company websites meet the structural citation-readiness threshold for both platforms.
of analysed websites pass the structural-readiness threshold for both platforms.
What leaders should see
AI exposure is concentrated. Website readiness is fragmented. A single composite score can hide a commercially important weakness: 33.2% of websites pass for GPT but fail for Gemini. Being prepared for the largest platform does not mean being prepared for the second-largest.
The practical priority is not to optimise equally for every platform. Start where market exposure and readiness risk overlap: GPT and Gemini first, with the largest remediation gap currently on Gemini.
Market exposure versus readiness
The chart compares each platform’s current model weight with the share of EU-27 websites scoring at least 70 for that platform.
| Platform | Platform weight | Mean readiness | Passing ≥70 | Pass rate |
|---|---|---|---|---|
| GPT | 54.6% | 69.1 | 32,317 | 50.6% |
| Gemini | 28.4% | 50.4 | 11,789 | 18.5% |
| Claude | 9.3% | 61.9 | 24,444 | 38.3% |
| Meta | 4.0% | 60.0 | 26,048 | 40.8% |
| Grok | 2.4% | 49.9 | 7,141 | 11.2% |
| Perplexity | 1.3% | 50.4 | 10,626 | 16.6% |
The GPT–Gemini readiness matrix
pass neither GPT nor Gemini.
pass GPT but fail Gemini — the largest hidden cross-platform gap.
pass both platforms.
pass Gemini but fail GPT.
Three management decisions
- Stop reading one composite score as universal readiness. Require platform-level reporting for GPT and Gemini.
- Prioritise the Gemini gap. Review Google-index visibility, valid structured data, organisation identity, visible dates, contact context and accessible content.
- Set a portfolio target. Measure the share of priority domains passing both platforms, not only the average score.
Methodology and evidence boundary
Website cohort
63,858 EU-27 company websites using the latest retained scan. The analysis requires Layer 1 crawler access above zero and a non-null structural readiness score for all six platforms.
The pass boundary is 70, matching the published VerisAI methodology.
Market-share input
The horizontal weights are the current worldwide platform weights used by the VerisAI citation-readiness model, with a reference date of 15 August 2026.
They are a separate market-intelligence input, not calculated from the 63,858 websites. See the full scoring methodology.
What the result does not prove
Structural readiness does not equal observed citation frequency, AI referral traffic or revenue. The analysis describes where market exposure and website prerequisites are misaligned. It does not claim causation, and it does not treat six platform points as a statistically robust correlation study.
Download the anonymised evidence
Every downloadable result is aggregated. No company names, domains, URLs, IP addresses, scan identifiers or raw page content are included.
Summary + methodology
JSON · cohort, definitions, limits and headline findings Platform results
CSV · weights, mean scores, counts and pass rates GPT × Gemini matrix
CSV · four aggregated readiness groups Reproducibility queries
SQL · aggregate-only analytical logic
JSON · cohort, definitions, limits and headline findings Platform results
CSV · weights, mean scores, counts and pass rates GPT × Gemini matrix
CSV · four aggregated readiness groups Reproducibility queries
SQL · aggregate-only analytical logic
Licence: CC BY 4.0. Attribution: VerisAI, EU AI Readiness Observatory, verisai.eu.