An AI visibility score is a composite metric, usually on a 0–100 scale, that summarizes how often and how prominently a brand appears in AI-generated answers across a set of prompts and engines. It typically combines mention rate, position, citation rate and sometimes sentiment, weighted by prompt importance. It's useful for trends and executive reporting, but should always be paired with its underlying metrics.
Executives want one number. AEO generates dozens: mention rates by platform, citation rates by topic, sentiment, position, share of voice. An AI visibility score condenses them into a single, trackable indicator of how visible your brand is in AI answers.
Used well, it focuses attention. Used carelessly, it hides what's really happening. Here's how to get it right.
What goes into an AI visibility score
Most scores combine some of these components:
| Component | What it captures |
|---|---|
| Mention rate | How often you appear at all |
| Position | How prominently you appear in lists |
| Citation rate | How often your content is used as a source |
| Recommendation rate | How often you're recommended for the user's need |
| Sentiment | Whether mentions are favorable |
| Platform coverage | Whether visibility is broad or concentrated |
| Prompt weighting | Whether you appear for the prompts that matter most |
Different tools weight these differently, so scores from different vendors aren't directly comparable. What matters is consistency over time within one methodology.
A simple scoring model you can build
For each prompt run on each engine:
- Presence: 1 if mentioned, 0 if not.
- Position weight: 1.0 if first, 0.8 second, 0.6 third, 0.4 for later.
- Citation bonus: +0.2 if your domain is cited.
- Sentiment modifier: ×1.0 positive or neutral, ×0.5 negative.
Run score = Presence × Position weight × Sentiment modifier + Citation bonus (capped at 1.0).
Then:
- Prompt score = average run score across runs and engines.
- Visibility score = weighted average of prompt scores (weights by business value), × 100.
This model is transparent, easy to explain and responsive to the actions you take.
How to use the score
For executive reporting
Show the score's trend over months, with competitors' scores for context. Pair it with one or two business metrics (AI referral conversions, branded search). See AEO reporting.
For goal setting
Set targets for the score in priority segments ("raise our shortlist-prompt visibility score from 32 to 45 this half").
For alerts
Sudden drops in the score flag issues worth investigating: blocked crawlers, a competitor's PR win, a model update.
How not to use it
- Don't optimize the score directly. Optimize the underlying drivers: sources, content, technical access.
- Don't compare across vendors. Methodologies differ.
- Don't hide the components. Always be able to drill down into mention rate, citations, sentiment and share of voice. See AI search performance metrics.
- Don't ignore sample size. Scores from a handful of runs are noisy. See LLM answer variability.
Visibility score vs share of voice
| AI visibility score | AI share of voice | |
|---|---|---|
| Measures | Your absolute presence and prominence | Your presence relative to competitors |
| Answers | "How visible are we?" | "Are we winning?" |
| Affected by competitors | Only indirectly | Directly |
| Best for | Trend tracking, goals | Competitive strategy |
Report both. A rising visibility score with falling share of voice means the whole category is getting more AI attention and competitors are gaining faster. See AI share of voice.
Choosing a tool
If you use an AI visibility platform, ask how its score is calculated, how many runs per prompt it uses, which engines it covers, and whether you can export the components. AEO-Rank reports visibility alongside position and sentiment for each prompt and platform, so you can always drill into the components behind the headline number. See the AI visibility tools buyer's guide.
A worked example
Imagine you track 40 prompts across ChatGPT, Gemini and Perplexity, running each prompt three times a month — 360 runs in total.
- Your brand appears in 144 runs: a 40% mention rate.
- When it appears, it is first in a third of runs, second in a third and third or later in the rest.
- Your domain is cited in 54 runs.
- Twelve mentions are negative.
Using the model above, the run scores average out to roughly 0.3 across all 360 runs, so your visibility score is about 30. Now weight the prompts: if your ten "best X for Y" prompts carry double weight and you are strongest on them, the weighted score rises; if you are weakest there, it falls. That is the point of weighting: the headline number should move when the prompts that matter move.
Next month, a new review-site listing lifts your mention rate on shortlist prompts and the score rises to 34. Drill down and you can see exactly why — which is what makes a transparent score useful.
Key takeaway: An AI visibility score is a helpful headline for trends and goals. Build it from transparent components, weight prompts by value, sample enough runs, and always pair it with share of voice and the underlying metrics.
