What isVisibility score
A visibility score is a single number, often on a 0-100 scale, that summarizes how prominently a brand appears in AI-generated answers across a defined set of prompts and engines.
It usually combines mention rate, position in the answer, recommendations and citations. It's useful for spotting trends, and meaningful only when you know the formula and method behind it.
Connected terms
The terms next to Visibility score in AI search, and where to read more.
AI share of voice→AI share of voice (AI SOV) is the share of all brand mentions in AI-generated answers that go to your brand, measured across a fixed set of prompts, engines and competitors.AI citation→An AI citation is a source an AI engine links or names to support its answer, such as a numbered footnote in Perplexity, a source link in ChatGPT search, or a link in Google AI Overviews.Monthly AI Visibility Report Template (With Looker Studio Dashboard)→A monthly AI visibility report template in Looker Studio: six sections, the metrics, a method box and how to explain changes, with MentionX data.All AI visibility tools guides→Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.
01What goes into one?
| Component | Typical role |
|---|---|
| Mention rate | Are you there at all? |
| Position | Named first or last? |
| Recommendation | Actively suggested? |
| Citation | Your URL used as a source? |
| Sentiment/accuracy | Described fairly and correctly? |
02Example
Illustrative only: a brand appears in 40% of answers, is recommended in 15% and cited in 10%, and usually sits mid-list. With weights of 0.4, 0.3, 0.2 and 0.1 for mention, recommendation, citation and position, its score would land somewhere in the 20s out of 100. The exact number matters less than how it moves month to month on the same prompts.
03What are the limits?
- Not comparable across tools that use different prompts and formulas.
- Hides causes: a stable score can mask a rise on one engine and a fall on another.
- Sensitive to prompt choice: adding branded prompts inflates it.
04How should you use it?
Use it as the headline in a report, always next to its components, the prompt count, the engines and the run dates. If it moves, the components should explain why.
05What questions should you ask a vendor about their score?
- Which components are included, and how are they weighted?
- Which engines and modes, in which country and language?
- How many runs per prompt, and how is variation handled?
- Can I see every raw answer behind the number?
- Has the formula changed, and when?
06From MentionX data: why one number hides a lotOriginal data
Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.
Any score built from these inputs inherits their spread. Our data shows how much each input moves on its own.
| Intent | Brand mentioned | Share of voice |
|---|---|---|
| Brand-specific | 100% | 64.3 |
| Comparison | 91.9% | 44.6 |
| Recommendation | 29.2% | 26.8 |
| Best-of | 26.6% | 22.7 |
| Problem | 5.7% | 7.7 |
Mention rates by engine ranged from 24.0% to 47.3%, and in 11 of 28 repeated prompt-engine pairs the brand was named in some runs and not others.
Our view: a visibility score is fine as a summary for executives if the components sit next to it. On its own, it can rise because the prompt mix changed.
Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.
07Worked example: Ledgerline's score drop
Ledgerline is a fictional accounts-payable tool. Figures are illustrative.
Ledgerline's vendor score falls from 52 to 44 in a month. The team breaks it down: mention rate is flat, citations are up, but the vendor added two problem-intent prompts to the default set, and those rarely name any brand.
Recalculated on the old prompt set, the score is 53. The team asks the vendor to lock the prompt set and report changes in a method note.
08What should you ask a vendor?
Ask which inputs the score uses and their weights; how many prompts and runs feed it; which engines, modes and locations; whether branded prompts are included; how competitors are chosen; and whether the method has changed since your last report. If a vendor can't answer, use the components instead.
Questions people ask
3 answersHow is a visibility score calculated?
+
It varies by tool. Most combine mention rate, position, recommendations and citations with weights. Ask for the formula.
Can I compare scores across tools?
+
No. Different prompts, engines and formulas make scores incomparable. Compare only within one method.
Should I report the score to leadership?
+
Report it alongside its components and method, so any change can be explained.