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

# Sentiment

> How an AI platform characterizes a brand when mentioning it — positive, neutral, negative, or mixed.

**Sentiment** in AI visibility measures how an AI platform characterizes a brand when it mentions the brand in a response. Cited classifies each mention into one of four categories: positive (favorable characterization), neutral (factual without evaluation), negative (unfavorable characterization), or mixed (both positive and negative in the same response).

## Why it matters

A high [mention rate](/glossary/mention-rate) with negative sentiment is worse than a lower mention rate with positive sentiment. When ChatGPT describes a brand as "overpriced" or "declining in quality," every mention works against the brand. Sentiment captures the qualitative dimension of AI visibility that quantitative metrics like mention rate and [share of voice](/glossary/share-of-voice) do not.

## How Cited uses it

Sentiment is classified per mention by the [Haiku parsing pipeline](/methodology/mention-extraction) and aggregated as a numeric score: (positive - negative) / total. Sentiment benchmarks are not yet published in the Cited Index because 86.5% of observations are classified as "mixed" — too concentrated for meaningful percentiles. See [Sentiment](/concepts/metrics/sentiment) for details.

## Related concepts

* [Sentiment (full page)](/concepts/metrics/sentiment) — why benchmarks are deferred
* [How we extract mentions](/methodology/mention-extraction) — the parsing pipeline
* [Brand mention](/glossary/brand-mention) — the atomic unit sentiment is classified on