The AI Visibility Measurement StackMetrics, Definitions, and How to Use Them
Five metrics for quantifying AI Visibility. Precise definitions, formulas, interpretation guidelines, and common misreadings -- organized as a standard reference.
Why Measurement Standards Matter
AI Visibility describes how present a brand is within AI-generated answers. But a concept alone does not drive improvement. Improvement requires measurement, and measurement requires well-defined metrics.
SEO became mainstream because it had clear metrics: rank, CTR, and traffic. Everyone measured the same things with the same vocabulary, making benchmarking and PDCA cycles possible.
AI Visibility does not yet have this standard. Through our own measurement practice and the 109-question GPR experiment, Sighted proposes the following five metrics as the "AI Visibility Measurement Stack."
1. Mention Rate
Definition: The percentage of AI responses that include your brand, company, or service name in the answer text
Formula: Responses containing mention / Total responses for target queries x 100
Interpretation Guidelines
| Query Type | Low | Medium | High |
|---|---|---|---|
| Brand queries | < 50% | 50-80% | > 80% |
| Industry queries | < 5% | 5-20% | > 20% |
Common Misreadings
A high Mention Rate does not automatically mean success. Negative mentions ("X has known issues with...") also count. Always pair with a qualitative sentiment check.
2. Citation Rate
Definition: The percentage of AI responses that cite your URL as a source
Formula: Responses citing your URL / Total responses for target queries x 100
Mention vs. Citation
Mention
"Sighted is a Japanese AEO tool" -- brand name appears in the text
Citation
"[Source: lp.sighted-aeo.com]" -- URL displayed as a source
Four State Patterns
| State | Mention | Citation | Meaning |
|---|---|---|---|
| Invisible | No | No | AI does not recognize the brand |
| Aware | Yes | No | Knows you exist but does not trust as a source |
| Trusted | Yes | Yes | Cited as a trusted source (ideal state) |
| Anonymous | No | Yes | URL cited but brand not named (rare) |
For a deeper diagnosis, see the Mention vs Citation diagnostic guide.
3. Response Share
Definition: The proportion of AI responses to a question set (Question Space) that mention or cite your brand
Formula: Responses mentioning/citing you / Total responses for the question set x 100
Response Share is the AI Visibility equivalent of "Share of Search" in SEO. Rather than tracking individual queries, it measures your aggregate presence across an entire question set.
For example, if your brand appears in 12 out of 30 responses for "AEO monitoring tool"-related queries, your Response Share is 40%. If competitor A appears in 15 (50%) and competitor B in 6 (20%), you rank second in this Question Space.
Common Misreadings
Response Share depends heavily on how the question set is defined. Narrow question sets yield higher shares; broader sets yield lower ones. Always use the same question set when comparing over time.
4. Placements
Definition: Where within the AI's answer your brand is positioned
Categories: Lead recommendation / Top of list / Mid-list / Footnote
Unlike the other metrics, Placements is a qualitative indicator that resists easy quantification. But its practical impact is significant.
Being introduced as "The top recommendation is X" at the start of a response versus appearing as "Other options include A, B, and C" at the end creates a massive difference in user perception and click-through behavior.
Placement Categories
Lead recommendation: Featured as the primary recommendation at the top of the answer
Top of list: Appears in position 1-2 of a recommendation list
Mid-list: Appears in position 3+ of a recommendation list
Footnote: Mentioned in supplementary or "see also" context
5. AEO Score (Composite Index)
Definition: A weighted composite of the above four metrics that represents overall AI Visibility health
Scale: 0-100
The AEO Score consolidates multiple metrics into a single number, making executive reporting and benchmark comparisons straightforward. It occupies a similar position to Domain Authority in SEO.
The detailed algorithm behind Sighted's AEO Score calculation is published in our Scoring Algorithm explainer.
Score Interpretation Guide
0-25
Invisible
26-50
Awareness
51-75
Citation
76-100
Recommendation
Combining the Five Metrics
The five metrics are most powerful when read as patterns, not in isolation.
High Mention + High Citation + Placement: Top
Ideal state. AI trusts your brand and recommends it prominently. Focus on maintenance.
High Mention + Low Citation
AI knows your brand but doesn't trust it as a source. Improve content quality, depth, and structure. See Content Quality Guide.
Low Response Share (all metrics low)
You do not exist in the AI's Question Space. Start with entity establishment and Question Space coverage. See AEO Execution Framework.
High Mention + Negative Sentiment
You are mentioned but in a negative context. Urgently identify the AI's sources and publish accurate, authoritative content to correct the narrative.
Measurement Cadence and Best Practices
Unlike SEO metrics that fluctuate in near-real-time, AI Visibility depends on AI model training cycles and content crawl frequency. Recommended measurement cadences:
| Metric | Frequency | Rationale |
|---|---|---|
| Mention Rate | Weekly | AI responses vary day-to-day; track weekly averages |
| Citation Rate | Weekly | Track changes after optimization interventions |
| Response Share | Monthly | Aggregate share shifts gradually |
| Placements | Monthly | Qualitative; periodic sampling is sufficient |
| AEO Score | Monthly | Composite score trend for executive reporting |
When starting measurement, allow a minimum 2-week baseline period with no optimization changes. This establishes your pure starting point. Details in the Monitoring Tools Guide.
Measure Your AI Visibility Score
See where you stand across all five metrics. Enter your domain to start.
Try free diagnosis