# How fixvis measures AI visibility

fixvis asks each AI engine the questions your buyers ask, reads every answer for your brand, and turns the result into one score with its margin of error.

## Six AI engines, one price

Every plan includes ChatGPT, Gemini, Claude, Perplexity, DeepSeek and Qwen, at the same price. There are no per-engine add-ons and no API keys to bring. fixvis also measures your visibility in Google and Bing search.

## How an answer is read

An answer can include you at three levels. fixvis records which one, because they are worth different amounts to you.

| Level | What it means | Example |
|---|---|---|
| **Mentioned** | Your brand appears in the answer | “…other options include Görüşme Yap.” |
| **Cited** | The engine links to your site as a source | A footnote to gorusmeyap.com |
| **Recommended** | The engine tells the buyer to choose you | “For therapists in Türkiye, Görüşme Yap is a strong fit.” |

## Sampled every day

AI answers vary from one run to the next. fixvis asks each question again on a daily schedule, so your score rests on many answers, not one. The free audit is a single pass; paid plans keep sampling.

## Every score has a range

Your AI visibility score runs from 0 to 100 and always shows its 95% range and the number of answers behind it. Two scans only disagree when their ranges do not overlap. fixvis marks a change as real only when the old and new ranges don't overlap.

> **Note:** In the app and glossary the method is called Share of Discoverability (SoD). On the website and in these docs it is the AI visibility score. They are the same number.