European AI Visibility Index — 2026-Q2

The first edition of the CEAVERS European AI Visibility Index covers 20 major European brands across five sectors, measured across six large language models in five European languages. The headline score of 51.5 out of 100 represents the mean brand visibility: roughly half of all category-relevant prompts resulted in a mention of the queried brand.

Key findings

Luxury and automotive brands lead AI visibility. LVMH (82), Volkswagen (79), and BMW (77) are the three most visible European brands in AI-generated responses globally. These brands benefit from a long digital paper trail, extensive Wikipedia coverage, and frequent cross-corroboration across independent news sources — all signals that raise citation probability in both parametric and retrieval-augmented LLMs.

Technology brands punch above their market capitalisation. SAP (71) and ASML (53) score significantly higher than their consumer-brand recognition would predict. Both companies appear frequently in technology journalism, policy discussions, and academic literature — domains that are systematically over-represented in LLM training corpora.

Financial services show the widest spread. The finance sector spans from BNP Paribas (58) to UniCredit (31), a 27-point range. This reflects unequal international coverage: French and German banks appear in English-language financial media far more than Italian or Spanish banks of equivalent size.

English-language scores average 12% higher than non-English. Across all 20 brands and all 6 LLMs, English-language prompts yield the highest mention rates. Portuguese-language prompts yield the lowest, averaging 10.2% below the cross-language mean. This gap is consistent with the documented English dominance in LLM training corpora.

Perplexity leads, Apple lags significantly. Among the six LLMs, Perplexity’s sonar-pro yields the highest brand visibility scores (mean +7.2% vs. cross-LLM average), benefiting from its real-time web retrieval architecture. Apple Intelligence scores the lowest (mean −17.9%), consistent with its newer model age and more conservative citation behaviour.

Score distribution

Platform comparison

Language effects

English prompts yield the highest visibility scores for all 20 brands in this panel. French (−1.0% vs. mean) and Spanish (−0.4%) are near parity. Italian (−4.8%) and Portuguese (−10.2%) show measurable under-representation, consistent with smaller-corpus effects documented in multilingual LLM benchmarks.

Brands headquartered in non-English-speaking countries do not show a systematic home-language advantage at this level of analysis. French brands (LVMH, BNP Paribas, Renault) do not score higher in French-language prompts than in English-language prompts, suggesting that the English corpus advantage outweighs geographic proximity effects.

Methodology summary

Scores are derived from a fixed bucket of 827 prompt templates per LLM × language combination (24,810 total evaluations), covering four query categories: direct brand queries (30%), category-level queries (40%), product/service recommendation queries (20%), and news/current-events queries (10%). A visibility score of 0–3 is assigned per evaluation (0 = not mentioned, 1 = passive mention, 2 = cited as example, 3 = cited as primary recommendation), normalized to a 0–100 scale. Full methodology at ceavers.org/methodology/.

Interpretation

A score of 100 would mean the brand appeared in every category-relevant AI response across all LLMs and languages. A score of 0 means no appearance in the test bucket. The headline mean of 51.5 is a panel average, not a single-brand score; individual brands range from 26 to 82.

These scores measure AI-response presence, not brand quality, consumer preference, or search-engine ranking. They reflect the current state of training data and retrieval architecture, both of which change over time.