METHODOLOGY / VERSION 1.0
Measured for change,
not certainty.
The AI Visibility Score™ is a normalized indicator of standardized, repeatable visibility. It is not an absolute ranking and does not predict every possible AI response.
Full audit score
Recommendation presence · 25%
Qualified recommendations ÷ eligible query-platform tests. A named list inclusion scores more than a passing mention.
Query coverage · 20%
Unique buyer-intent queries with a qualified brand mention ÷ eligible queries tested.
Competitive share of voice · 15%
Weighted brand mentions ÷ all weighted mentions among the locked comparison set.
Entity accuracy · 15%
Correct facts ÷ checkable facts across name, category, services, audience, location and positioning.
Citation quality · 10%
Attributable citations weighted by relevance, independence and source authority.
Platform diversity · 10%
Supported platforms producing qualified visibility, normalized by platforms successfully tested.
Geographic consistency · 5%
Correct market relevance across localized discovery queries.
Query-universe rules
Each business receives a locked universe across eight intent families: broad discovery, local discovery, comparison, recommendation, expertise, problem-solving, transactional intent and competitor alternatives. Queries must be natural, non-branded where discovery is being tested, relevant to a purchase decision, and reproducible. The same universe is reused for trend measurement; changes are versioned.
Competitor rules
The comparison set combines customer-named competitors, locally relevant operators and brands observed repeatedly during a calibration run. Inclusion requires category and market relevance. The final set is frozen for a reporting cycle; emerging competitors are reported separately until the next rebasing.
Confidence
High: at least 90% of planned query-platform tests completed with a stable comparison set. Moderate: 70–89% completed or one material input is incomplete. Low: below 70%; directional findings may be shown, but a headline score should not be compared with prior periods.
Score bands
0–25 Low visibility · 26–50 Emerging visibility · 51–70 Competitive visibility · 71–85 Strong visibility · 86–100 Category authority.
Limitations
Responses vary by model version, geography, personalization, query wording and time. Some platforms restrict automated testing. Tests must respect platform terms and technical limits. Missing observations reduce confidence and are never disguised as zero. No score guarantees citation, recommendation, traffic or revenue.