What is an AI visibility score? Definition and how it is built

An AI visibility score rolls mention rate, sentiment, share of voice and position into one number. Learn how the composite is built and what it can hide.

  • Alvaro Peña de Luna Alvaro Peña de Luna
  • date icon

    Wednesday, Sep 30, 2026

An AI visibility score is a composite number, usually on a 0 to 100 scale, that summarizes how visible a brand is in AI-generated answers by combining several underlying metrics, such as how often the brand is named, how favorably it is described and how it compares with competitors.

There is no standard formula. Each tool that shows one picks its own inputs and weights, so the score only makes sense once you know what goes into it.

How an AI visibility score works

A score starts from a set of prompts run on one or more AI engines, such as ChatGPT, Perplexity, Google AI Overviews and Google AI Mode. Each answer is read for the brands it names, how it names them and which sources it cites. Those observations become a handful of metrics, and the score is a weighted blend of them, rescaled to 0 to 100.

Three decisions shape the result:

  • Inputs: which metrics go into the score and which stay out.
  • Weights: how much each metric counts towards the total.
  • Normalization: how a percentage, a 0 to 100 index and a position (where lower is better) are put on the same scale so they can be added together.

Change any of the three and the same brand, with the same answers, gets a different score.

The metrics behind the score

Most AI visibility scores draw on some combination of these metrics:

  • Mention rate: the share of answers that name your brand. See mention rate for the full definition.
  • Sentiment: whether the answers talk about you in a positive, neutral or negative way.
  • Share of voice: your presence in the answers compared with your competitors'.
  • Position: how early in the answer you are named, or how high your pages sit among the cited sources. Lower is better, so a score has to invert it before adding it.
  • Citations: how often the engine links to your site as a source. See AI citations.

Example: how the components combine

Take a hypothetical score that blends three inputs already on a 0 to 100 scale:

AI visibility score = 0.4 × mention rate + 0.3 × share of voice + 0.3 × favorability

Brand A is named in 30 of 100 answers (mention rate 30), holds a 40% share of voice and has a favorability of 75. Its score is 0.4 × 30 + 0.3 × 40 + 0.3 × 75 = 12 + 12 + 22.5 = 46.5.

Brand B is named in 60 of 100 answers (mention rate 60), holds a 20% share of voice and has a favorability of 40. Its score is 24 + 6 + 12 = 42.

The two scores sit less than five points apart, yet the brands have opposite problems. Brand A is rarely named but well regarded when it is. Brand B is named often, but competitors outweigh it and the tone is lukewarm. The single number hides the fix each one needs, which is why you should always read the components behind it.

A score works as a headline. It tells a stakeholder at a glance whether AI visibility is rising or falling, and it is handy for comparing groups of prompts inside the same tool, for example product queries against comparison queries, because the method stays constant.

What it cannot do is tell you what to change. Decisions come from the components: a low mention rate points to gaps in your content and third-party coverage, weak sentiment points to how sources describe you, and a low share of voice points to specific competitors winning the same answers. For the full set of metrics and how to read them, see our guide to GEO metrics.

Common mistakes

  • Comparing scores from different tools. Different inputs, weights and prompt sets produce different numbers for the same brand.
  • Trusting a single check. AI answers vary between runs. A score from one run is mostly noise; track on a regular cadence and watch the trend.
  • Ignoring the prompt set. A score built on prompts that already contain your brand name will look high and mean little. Unbranded category prompts show whether engines choose you.
  • Blending engines. Your visibility on ChatGPT and on Google AI Overviews can differ sharply, and one averaged number hides the gap.
  • Optimizing the score instead of the answers. The goal is to be named, cited and recommended, not to move a number whose recipe you do not control.

How Mencoro reports the components

Mencoro does not collapse your data into one visibility score. It reports each component on its own, per engine and country, across ChatGPT, Perplexity, Claude, Google AI Overview and Google AI Mode, so you can see which one moved:

  • Coverage: checks in which you were named ÷ checks that returned an answer × 100. A link to your site alone does not count.
  • Favorability: (positive + ½ × neutral) ÷ all mentions × 100, where 50 means your mentions balance out to neutral.
  • Share of voice: your weighted mentions against your tracked competitors', where each mention weighs type × tone × qualifier. A positive, direct recommendation weighs 1.0; a neutral, direct reference weighs 0.14.
  • Mention and link position: how early you are named and how high your pages sit among the sources, on the days you appear.

If you need a single figure for a report, you can build your own blend from these, as long as the components stay next to it. The methodology page explains every formula, the visibility feature shows how the metrics appear in the app, and the free AI visibility checker gives you a first sample of whether AI engines name you next to your competitors.

FAQ

Frequently asked questions

There is no universal benchmark, because every tool builds its score differently and the result depends on your market and your prompts. Compare your score with competitors measured by the same tool on the same prompts, and focus on the trend over time rather than on one figure.
No. Each tool chooses its own metrics, weights, engines and prompt sets, so the same brand can get very different scores in two tools. Only compare scores produced with the same method.
Look at the components before you react. A drop can come from being named less often, from a shift in tone, from a competitor gaining ground or from a change in the prompt set. AI answers also vary from run to run, so confirm the drop holds across several checks.
No. Mencoro reports Coverage, Favorability, Share of voice and mention and link positions separately, per engine and country, so you can see which one moved. If you need one number for a report, you can blend them yourself and keep the components next to it.

Start tracking your brand in AI search today

Monitor how AI engines cite your brand, track keyword positions, and benchmark against competitors, all in one platform.