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

# Sentiment

> How AI describes your brand: positive, neutral or negative, per engine.

Sentiment shows how AI describes your brand when it mentions you. It appears on [AI Visibility](/analytics/visibility/ai-visibility) as **How AI describes *your brand***.

## How it's judged

After each scan, RankRaft AI reviews the opening of every answer, where an engine usually states its verdict, and labels its tone toward each brand it names as **Positive**, **Neutral** or **Negative**.

Neutral is the correct label for a plain factual listing. Most answers that simply list options are neutral, so a high neutral share is normal and not a problem.

## How it's shown

* Split per engine.
* Only answers that **mention you** count.
* The legend shows **Positive**, **Neutral**, **Negative** and **Mixed or unknown**. Answers that couldn't be judged count as neutral, so each bar adds up to 100%.

## What AI says about you

The **What AI says about *your brand*** card lists the two-word phrases that keep recurring around your name in answers, such as *authentic taste* or *expensive shipping*, with how often each appears. The phrases share the overall tone of the answers they came from. Use them to see which associations AI has formed, and whether they match how you want to be known.

> [!TIP]
>
> If a negative phrase keeps recurring, search for it in [Prompts & Answers](/analytics/prompts/prompts-and-answers) and check the **Sources cited** under those answers. The phrase usually comes from one or two specific pages.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.