> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fixaeo.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Visibility Score

> The share of AI answers that mention your brand.

Visibility Score is the number FixAEO leads with, and the one worth watching daily.

Visibility Score is the percentage of AI-generated answers that mention your brand. If FixAEO captured 120 answers about your category yesterday and you were named in 42 of them, your Visibility Score is 35%.

## How is Visibility Score calculated?

FixAEO captures one answer per prompt, per engine, per region, per day. Each answer is checked for your brand name. The score is a plain ratio of those two counts.

```
Visibility Score = brand mentions ÷ answers captured × 100

Per engine:
engine score = mentions on that engine ÷ answers captured on that engine × 100
```

The change beside the score is reported in percentage points, not percent.

<Warning>
  Visibility Score counts mentions, not links. An answer that recommends you in prose without linking to you still counts. For link-level data, see [Citations](/citations).
</Warning>

## What's a good Visibility Score?

There's no universal benchmark, and you should distrust anyone who hands you one. Your prompt list sets the denominator, so two brands with different prompt lists aren't comparable. Five branded questions will score far higher than fifty category questions, and mean far less.

Two comparisons do work. The first is your own trend over 30 days. The second is your rank against rivals in [Industry benchmarks](/industry), where everyone shares your denominator.

## How do you improve it?

Start with the per-engine split on [Model](/model). A weak overall score is usually one or two engines, not all of them. Check the engine rail on your [dashboard](/dashboard) first, so an outage doesn't read as a drop.

Then find the themes dragging it down. [Topics and labels](/topics-and-labels) groups your prompts so you can see where you're never named. [Sources](/sources) shows which sites the engines read before answering. [Improve](/improve) turns that into specific work.

If the score looks wrong rather than low, your prompt list is usually the cause. See [Setting up your prompts](/setting-up-prompts).

<CardGroup cols={2}>
  <Card icon="chart-pie" title="Share of voice" href="/metrics/share-of-voice">
    The same mentions, measured against your competitors instead of against answers.
  </Card>

  <Card icon="ranking-star" title="Position" href="/metrics/position">
    Not just whether you're named, but how early.
  </Card>

  <Card icon="face-smile" title="Sentiment" href="/metrics/sentiment">
    How the answer describes you once it names you.
  </Card>

  <Card icon="gauge" title="Metrics overview" href="/metrics-overview">
    All four metrics and how they fit together.
  </Card>
</CardGroup>