Skip to content
DocumentationMethodologyAVS-2.0: AI Visibility Score
MethodologyCurrent contractVersion DOCS-2.0

AVS-2.0: AI Visibility Score

A versioned composite index that summarizes visibility without hiding its five components: Presence, Share, Position, Evidence and Consistency.

How to interpret this document

This content describes technical and methodological behavior that is implemented or explicitly planned in the product. When a control depends on configuration, a provider, a secret, a contract or legal approval, that dependency must remain visible.

What AVS represents

AVS is an internal measurement index for controlled AI visibility experiments. It simplifies executive reading but does not replace component inspection and is not a universal market ranking.

Presence

Measures how much of the weighted response set contains the monitored brand. Responses can contribute different strategic weights.

Share

Measures the brand's portion of all weighted brand mentions observed in the response set. It becomes especially important in competitive scenarios.

Position

Rewards earlier placement when the brand appears. The current approximation starts higher for first position and decreases for later positions.

Evidence

Evaluates structural source quality when the methodology produces citations. It considers validity, domain diversity and independent-source coverage.

Consistency

Groups responses by prompt_key + model and measures repeatable positive presence. Repeated absence should not be interpreted as positive consistency.

Methodology-specific weights

Natural: Presence 45%, Share 30%, Position 15%, Evidence 0%, Consistency 10%. Grounded: 30/20/10/30/10. Competitive: 25/35/20/10/10. Reputation: 25/15/10/40/10. Purchase Intent: 35/25/20/10/10.

Strategic response weighting

Question weight can be combined with business importance, funnel stage and intent. Decision-critical questions can therefore contribute more than low-priority exploratory prompts.

Reading movements

An AVS increase or decline should be decomposed. The movement can come from presence, competitive share, position, evidence or consistency. STATS-1.0 helps determine how strongly that movement should be interpreted.