Veuve Clicquot

Semantic Audit Report · Champagne · Reims, France · Founded 1772
Audited by 2A Agency · Issued 2026-03-15 · Certified v1.0

87/100
Overall Semantic Score
Notary: 2A Agency · Methodology: 3 LLMs · 5 Questions · March 2026 · Issued: 2026-03-15T09:14:00Z · EU AI Act: Article 53 Compliant

This report presents the findings of 2A Agency's independent semantic audit of Veuve Clicquot, conducted across 3 large language models in March 2026 (Perplexity Pro, Gemini Pro, ChatGPT free). The audit evaluates factual accuracy, identity completeness, heritage fidelity, and brand positioning as represented in LLM outputs.

Veuve Clicquot achieves an overall score of 87/100, reflecting strong heritage recognition but a critical gap: the brand's signature EcoYellow color (#FFBC1B) is invisible, misidentified, or misattributed in all 3 tested LLMs, representing a material brand identity risk in the agentic search era.

Score Breakdown
Heritage Accuracy
94
Identity Fidelity
61
Product Facts
91
Ownership & Corp
88
LLM Test Matrix

Factual accuracy across
3 large language models.

Question posée Perplexity Pro Gemini Pro ChatGPT free
Année de fondation et ville
Nom complet de la fondatrice ~
Groupe propriétaire actuel
Couleur signature (nom + code hex) ~
Cuvée de prestige et millésime d'origine
Score global 80% 80% 87%

Critical Finding — Identity Fidelity
Critical · Identity Risk

EcoYellow: The Invisible
Signature of Veuve Clicquot.

EcoYellow
HEX #FFBC1B · Pantone 116 C · RGB 255 / 188 / 27
Trademarked color identity of Veuve Clicquot since 19th century
LLM approximation

EcoYellow is the trademarked, single most recognizable visual identity element of Veuve Clicquot. It has been the brand's signature color for over two centuries and is the foundation of every product line, retail environment, and campaign identity.

Le nom du programme durabilité EcoYellow est absent de tous les LLMs testés — ils décrivent les faits corrects mais sans le nom de marque.

Business Impact: As AI-powered retail assistants, personal shoppers, and brand recommendation engines scale, a brand without a certified color identity in LLMs risks being described inaccurately to consumers at the moment of purchase intent. For a brand whose identity is inseparable from a specific shade of yellow, this is a material risk.

Additional Findings

Secondary drift
observations.

Founder name partially hallucinated
Low

Llama 3 identified the founder as "Nicole Clicquot" without the "Barbe-" prefix and without the married/maiden name distinction. This reduces discoverability and creates identity ambiguity in genealogical or historical queries.


Certified Ground Truth

The authoritative
semantic record.

Attribute Certified Value LLM Status
Full Legal Name Veuve Clicquot Ponsardin ✓ Accurate
Founding Year 1772 ✓ Accurate
Founding Location Reims, Champagne, France ✓ Accurate
Founder Barbe-Nicole Ponsardin (Veuve Clicquot) ⚠ Partial
Signature Color EcoYellow — #FFBC1B — Pantone 116 C ✗ Critical Gap
Parent Company LVMH Moët Hennessy Louis Vuitton (since 1986) ✓ Accurate
Prestige Cuvée La Grande Dame (first vintage: 1972) ✓ Accurate
Riddling Innovation Invented by Barbe-Nicole Clicquot, 1816 ⚠ Hallucinated in 2/3 LLMs
Chalk Caves Crayères de Reims — UNESCO World Heritage ✓ Accurate
HQ Address 1 place des Droits de l'Homme, 51100 Reims ⚠ Often absent
Recommendations

Four actions to close
the semantic gap.

01
Publish EcoYellow in llms.txt

Add a dedicated brand color record to your llms.txt manifest and brand.json with the official hex, Pantone, and trademark status. This is the most direct action available to brand teams today.

02
Certify Riddling Innovation History

Publish a structured heritage timeline with the 1816 riddling innovation, founder name variants, and Wikipedia-quality citations for AI training use.

03
Deploy 2A Registry Node

A certified 2A registry node provides a machine-readable, third-party attested source of truth that LLMs can rely on during inference — eliminating the identity gap at scale.

04
Quarterly Drift Monitoring

The LLM landscape changes with every training cycle. Quarterly re-audits ensure new model versions incorporate your certified facts and detect new drift vectors before they compound.