Skip to main content
Sentiment Score is the average tone of AI mentions of your brand on a 0 to 100 scale: (sum of per-mention sentiment scores ÷ total mentions). Erlin runs LLM-based analysis on the language around every brand mention and assigns a score: 0 for negative, 50 for neutral, 100 for positive. It is the AI-search version of brand health tracking.

​
Why it matters

A brand can be mentioned often and still lose deals if AI frames it negatively. “Users report slow support and confusing pricing” drives buyers away as effectively as not being mentioned. Sentiment Score catches tone problems early. A drop from 75 to 55 over two weeks is a signal that something changed in the public sources AI draws from, and you have a window to investigate before more users see it.

​
What Sentiment Score tells you

Sentiment Score measures tone. AI describing your brand as “highly recommended for fast-growing teams” produces a score near 100. AI saying “users report long support waits” produces a score near 0. The dashboard averages every mention into a single number you can track week over week. The most useful signal is the change over time. A stable score means nothing new is influencing AI’s language about you. A sharp move means a specific source surfaced and shifted how AI talks about your brand.

​
How Erlin calculates it

For every AI response that mentions your brand, Erlin sends the language around the mention to an LLM-based sentiment classifier. Each mention receives a score:
  • Positive = 100
  • Neutral = 50
  • Negative = 0
The dashboard score is the average:
Worked example: in the last 30 days you have 50 mentions. Of those: 30 positive, 18 neutral, 2 negative. (30 × 100 + 18 × 50 + 2 × 0) ÷ 50 = 78

​
How to read it in Erlin

A stable Sentiment Score above 70 means AI consistently speaks well of your brand. A falling score means new negative language is entering AI’s responses. Compare platforms: if Sentiment is 80 on ChatGPT but 45 on Perplexity, Perplexity may be pulling from a source with critical content that ChatGPT does not weight as heavily. To see the specific language AI used, open the prompt detail view in Prompt tracking and read the full response excerpt for any individual mention.

​
How to improve it

Close visible complaints. Negative sentiment is driven by support issues, billing confusion, and product problems that show up on review sites and forums. AI picks these up and quotes them. Fix the underlying issue. Once fresh content no longer surfaces complaints, sentiment recovers over time. Surface strong reviews. If you have enthusiastic customers on G2, Capterra, or industry forums, make sure those reviews are visible and well-cited. Pages with strongly positive reviews that carry high citation coverage lift Sentiment directly. Investigate every sudden drop. A sharp negative shift almost always traces to a specific source: a critical review article, a Reddit thread, a competitor comparison, or a news article about an outage. The Citations module shows which sources appeared recently. Cross-reference with the timing of the Sentiment drop to find the trigger.

Visibility Score

How often AI mentions your brand across tracked prompts.

Citation Rate

Your share of the sources AI cites in its answers.

Share of Voice

When AI lists products, how big a slice your brand takes.

Avg Position

Where your brand lands when AI ranks recommendations.