Sentiment Score
A measure of whether an AI's mention of a brand reads as positive, neutral, or negative.
What is Sentiment Score?
Sentiment score captures the emotional tone of how a brand is described (positive, neutral, or negative) rather than simply whether it's mentioned at all. For instance, an AI assistant calling a product "a reliable, well-reviewed choice" would be tagged positive, while a purely factual listing with no evaluative language would be tagged neutral.
Why it matters
The technique draws on sentiment analysis (also called opinion mining), a long-established NLP discipline that classifies text polarity using lexicon-based rules, trained classifiers, or, increasingly, an LLM prompted to judge a passage's tone. When applied to AI-answer monitoring, each brand mention is converted into a category or numeric scale and then averaged across many tracked prompts and AI models to produce an aggregate score; there is no universal scale or threshold across the industry. Because the underlying classifier's own judgment does the scoring (often another LLM), results can differ meaningfully between tools, and borderline phrasing, such as a brand described only as "cheaper" than a competitor, may be classified inconsistently. Ranketta computes a sentiment score for tracked brand mentions this way, using an LLM classifier to judge tone across every tracked prompt and AI provider.