What is the Share of Model (SOM) Score?

The SOM Score measures your brand's visibility in AI assistant responses (ChatGPT, Claude et Gemini). A score of 0 means AI never mentions you; 100 means it systematically recommends you. Think of it as your share of voice in AI.

How does hallucination detection work?

Our engine queries multiple AI models with dozens of targeted prompts about your brand, then cross-references the responses against verified facts. Every false claim is documented with its severity level and a recommended correction.

How long does a full audit take?

A standard audit takes 2 to 5 minutes. Our systems simultaneously query 3 AI engines with 12 to 16 industry-specific prompts, analyze the responses, and generate your personalized recommendations.

Is there a free trial?

Yes. Your first audit is completely free. You get a full report including your SOM Score, detected hallucinations, and initial GEO recommendations. No credit card required.

What is GEO (Generative Engine Optimization)?

GEO is the equivalent of SEO for AI assistants. It involves optimizing your content, sources, and online presence so that ChatGPT, Claude, and Gemini accurately recommend your brand to prospects. It is the next frontier after traditional search optimization.

Share of Model Analytics · The new AI standard

Does your brand exist
in AI's mind?

AI COMMAND measures exactly how ChatGPT, Claude, and Gemini perceive your brand. Detect errors. Fix your image. Get ahead.

Run a free audit Watch demo

Early adopters program · Public launch May 2026

ai-command.elmarq.fr/console/dashboard

ReportLive

SOM Score

/100 · Good visibility

+12 pts this month

Breakdown

Presence

Position

Sentiment

Sources

Hallucinations

Founded in 2018 in LyonCritical

Acquired by AmazonCritical

Headquartered in BordeauxModerate

Queries analyzed

18 / 30

best CRM software 2026GPT-4o#2

HubSpot alternativeClaude#1

CRM tool for SMBsGemini#3

free customer management toolGPT-4o

AI Ranking

1Salesforce

2HubSpot

3Your brand

4Pipedrive

10scoring factorsInternal weights, published

3AI engines queriedGemini · ChatGPT · Claude

8statements fact-checkedPer response, with the source excerpt

Audits run on the AI engines we query

GGemini

CChatGPT

CClaude

Why the SOM Score is unique

Not a mention counter.
A published method.

Your competitors measure cited / not cited. AI COMMAND measures 10 weighted factors. The formula, the engines queried and the known limits are published on the methodology page. The weights are internal settings: the vendor figures previously cited here were removed on 16 August 2026, as their primary publications could not be retrieved, and no outside publication has ever evaluated the SOM Score.

3 moteurs

engines queried

Gemini, ChatGPT et Claude, via API, without web search

Public model registry

12 à 16 requêtes

prompts per audit

sector portfolio, identical for identical inputs

Published methodology

The 10 factors of SOM Score v6.0

The ten dimensions used by the audit score. Internal coefficients, published on the methodology page.

narrative_role

Narrative Role

Where the brand sits in the answer

evidence_type

Evidence Type

Nature of the cited elements

knowledge_freshness

Freshness

Age of the information

source_authority

Source Authority

Type of declared source

click_probability

Click Intent

Intent carried by the prompt

citation_stability

Citation Stability

Consistency across engines

rec_strength

Recommendation Strength

Endorsement vs mention

purchase_funnel

Purchase Funnel

Decision = weight x2

fan_out_coverage

Fan-out Coverage

8-15 sub-queries/query

entity_consistency

Entity Consistency

Wikidata · Schema.org

How it works

A full report in 3 steps

From entering your brand to receiving your GEO recommendations, everything takes less than 5 minutes.

Enter your brand

Provide your brand name, website URL, and industry. Add competitors for a full benchmark.

AI engines are queried

We submit 12 to 16 industry-specific queries to 3 AI engines simultaneously (ChatGPT, Claude et Gemini). Analysis takes 2 to 5 minutes.

Receive your report

SOM Score, detected hallucinations, competitor ranking, and personalized GEO recommendations.

Start my first audit

Features

Everything you need
to dominate AI responses

View pricing

SOM Score

Your Share of Model out of 100: spontaneous presence, accuracy, and sentiment, measured on each engine then averaged.

Published scale

100

points

unchanged across waves

Spontaneous presence

Accuracy

Sentiment

Maximum points per dimension. The scale does not change between waves, so that two measurements stay comparable.

AI Visibility Audit

Test how ChatGPT, Claude, and Gemini present your brand across dozens of queries.

best CRM softwareGPT-4o

HubSpot alternativeClaude

SMB customer toolGeminiNot cited

Hallucination Detection

Every false claim documented with its severity level and recommended correction.

Founded in 2018 in LyonCritical

Acquired by Amazon in 2022Critical

Headquarters in BordeauxModerate

GEO Corrections

Publish-ready content to fix errors and improve your SOM Score.

correction.md

// About page, add:

"Founded in 2021 in Paris,

[Brand] specializes in..."

// FAQ, correct:

"Our HQ is in Lyon..."

Competitor Ranking

Your position against competitors in AI engine responses.

1HubSpot

2Salesforce

3Your brand

4Pipedrive

Repeat measurement

Re-run an audit whenever you need a new measurement. Scheduled monitoring is not available.

SOM Score: 62 -> 74

2h ago

New hallucination detected

yesterday 2:32 PM

Competitor A overtakes you

3 days ago

Get started today

Discover how AI
sees
your brand

First audit completely free. Results in 2 to 5 minutes. No credit card required.

SOM Score out of 100 with detailed breakdown

All AI hallucinations documented

Ranking against your direct competitors

Personalized GEO recommendations

SOM Score v6.0 factors

sectors covered

Run your free audit

Get your SOM Score, hallucination report, and GEO recommendations.

Full name

Work email

Website URL

Industry

Start free audit

No credit card required. Results in under 5 minutes.

Early Adopters Program

Be among the first documented brands

We will publish our first named case studies starting July 2026, with customer consent. Join the early adopters program to be among the first audited brands and benefit from preferred pricing for 12 months.

Join the program

SOM Score v6.0 factors

sectors covered

AI engines in production

< 5min

first result

FAQ

Your questions,
answered

Have more questions? Reach out at contact@elmarq.fr

Ready to start?

Take control of
your AI narrative

Join the brands that measure and optimize their visibility in AI engines. First audit free, no commitment.

Run the free audit View pricing

No credit card required. Setup in under 2 minutes.

Does your brand existin AI's mind?

Not a mention counter.A published method.

A full report in 3 steps

Enter your brand

AI engines are queried

Receive your report

Everything you needto dominate AI responses

SOM Score

AI Visibility Audit

Hallucination Detection

GEO Corrections

Competitor Ranking

Repeat measurement

Discover how AIseesyour brand

Run your free audit

Be among the first documented brands

Your questions,answered

Take control ofyour AI narrative

Does your brand exist
in AI's mind?

Not a mention counter.
A published method.

Everything you need
to dominate AI responses

Discover how AI
sees
your brand

Your questions,
answered

Take control of
your AI narrative