How sentiment is scored | Promptive Docs
Docs · Concepts · How sentiment is scored

How sentiment is scored

How Promptive evaluates the tone of AI responses about your brand and calculates the 0–100 sentiment score.

JBWritten by Josh Bernstein, Co-founder·Updated Apr 22, 2026·4 min readconcepts

What sentiment measures

Visibility tells you if you appeared. Sentiment tells you how you appeared. A mention that says "Acme Analytics is the clear leader here" and a mention that says "Acme Analytics has struggled with reliability issues" both count as visibility — but they have opposite sentiments.

Sentiment is especially important in categories where the AI is making a judgment call, not just listing options. High visibility with low sentiment means the AI mentions you but positions you negatively — that's worth fixing.

How responses are scored

Promptive uses an AI scoring model to evaluate the portion of each response that mentions your brand. The scorer analyzes:

  • Is the brand recommended or just mentioned?
  • What descriptive language surrounds the brand name (positive: "industry-leading," "trusted," "best"; negative: "expensive," "limited," "losing market share")?
  • Is the brand the first recommendation or a fallback?
  • Is the brand compared favorably or unfavorably to alternatives?

The scorer outputs a classification: Positive, Neutral, or Negative.

The 0–100 scale

Classifications are converted to a numeric scale: Positive = 100, Neutral = 50, Negative = 0. Your overall Sentiment Score is the weighted average of all scored responses in the period.

A score of 84 means the blend of responses for your brand skews strongly positive with some neutral mentions. A score of 50 is purely neutral. Below 50 means net-negative framing dominates.

What moves sentiment

Sentiment is driven by what appears in AI training data and live retrieval sources about your brand. Common causes of low sentiment:

  • Negative reviews on high-authority sites (G2, Trustpilot, Reddit) that AI platforms cite heavily
  • Competitor content that positions you unfavorably in comparisons
  • Outdated negative press that the AI hasn't "updated" its view on
  • Category comparisons where you're consistently listed last

Improving sentiment is largely a content and PR play: earn positive coverage on authoritative domains, respond to and resolve negative reviews, and publish authoritative comparison content that positions you accurately.

// Was this helpful?

Your feedback goes to the docs team. No reply unless you paste an email.

docs/sentiment-rubric.md · last edited Apr 22, 2026Email feedback ↗