Sentiment

See how AI describes your brand when it mentions it, then open the answers behind the tone.

Sentiment overview

Open AI visibility → Sentiment. Use the category and model filters to focus the results.

This screen uses the latest completed answers available for each prompt and AI model. It includes prompts that name a brand. It does not offer a selectable 7-, 30- or 90-day period. To compare successive measurements, use AI trends.

Sentiment is measured around mentions of your brand. An answer without a scored mention is not a neutral verdict.

Read the distribution and model breakdown

Sentiment distribution groups scored answers into five tones, from very negative through neutral to very positive. Check the counts as well as the shape of the distribution.

By AI platform shows average sentiment and the number of runs behind it for each AI model. A model with no scored mentions has no sentiment result to compare.

The score summarises brand-related tone. Individual phrases in an answer may have different tones, so inspect the answer when you need to understand a value.

Compare categories

By category shows the sentiment mix for each category. A category without mentions has an empty/no-mentions state rather than a neutral score. Unassigned prompts appear separately when they have scored mentions.

Keep the same category and model filters when comparing the distribution and model breakdown. Counts can change when you change that selection.

Open positive and negative answers

Most positive prompts and Most negative prompts list prompts by average sentiment in their latest available answers.

Click a prompt to open its detail page, then select the model and sample you want to read. Inspect the highlighted phrases around your brand and the cited pages alongside the answer. A summary score alone does not identify which sentence or source needs attention.

For the detail controls, see Prompts & answers. For cited pages across prompts, see Sources.