Reading How AI Engines Talk About Your Brand
Key Concepts
Sentiment Prompts: A separate set of prompts, distinct from visibility prompts, that directly ask AI engines to evaluate a specific company on a specific topic. These are auto-generated based on your configuration (number of topics × number of competitors). They are not inferred from your visibility data.
Perception vs. Reality Problem: When a negative theme appears in sentiment data, the cited sources tell you whether it's a content problem (a single article or outdated review driving the narrative) or a product problem (the engines are accurately reflecting a real limitation). The fix is fundamentally different for each.
Quick Reference
Sentiment View Overview
The sentiment view shows:
- Overall positive sentiment % and its weekly trend
- Positive themes with individual mention counts and trend direction
- Negative themes with individual mention counts and trend direction
Investigating a Theme
When a theme catches your attention (especially negative themes with rising mention counts):
- Click into the theme to see the full engine evaluation
- Read the structured breakdown the engine provided (strengths, limitations, positioning)
- Scroll to the citations at the bottom of the response
- Identify what sources are feeding the narrative
What to Do With What You Find
- One article or outdated source driving a negative theme: Create newer content to reshape the narrative (reviews, case studies, customer stories)
- Multiple sources accurately reflecting a real issue: The fix is product-side, not content-side (documentation, onboarding, the product itself)
- Positive theme trending up: Note it, reinforce it
- Negative theme trending down: Something is already working - identify what changed
Reading the Trends
A single negative mention is noise. A negative theme growing over multiple weeks is signal worth investigating. Focus your attention on direction, not snapshot counts.
