What it is
The Sentiment tab analyzes the emotional and contextual language used when AI engines mention your brand and competitors.
It shows:
Whether brands are described positively, neutrally, or negatively
How sentiment differs by source (Web Grounding vs Training Data)
Which keywords drive perception
How sentiment trends over time
This is your brand perception dashboard for AI search.
Why it matters
AI-generated responses don’t just rank brands, they describe them.
The tone of that description affects:
Perceived authority
Trust
Purchase intent
Competitive positioning
Two brands may rank similarly but be framed very differently.
Sentiment analysis helps you understand:
How AI models position your brand
Whether perception is improving or declining
Which keywords shape your narrative
Filters & controls
The same filters and controls used on other dashboard tabs apply here.
You can filter by:
Filter by:
AI Engine: View performance across all engines or isolate one (e.g. Gemini, ChatGPT, Perplexity).
Topic: Focus on a specific topic group.
Tags: Filter by custom term tags.
Queries: Narrow down to specific tracked search terms.
All charts and tables update dynamically based on selected filters.
Time & aggregation controls
You can also adjust:
Timeframe: 24h, 7d, 30d, 3m, 1y
Aggregation: Hourly, Daily, Weekly, Monthly
Metric type: Average, Latest
These settings control how metrics are calculated and displayed.
Sentiment analysis by brand (main table)
The main table lists all detected brands and their sentiment metrics.
For each brand, you can see:
Source (Web Grounding / Training Data)
Sentiment score
Total positive keywords
Total neutral keywords
Total negative keywords
Total mentions
What “Source” means
Sentiment can come from two types of AI data:
Web Grounding (GR)
Real-time web content referenced by the AI engine
Sentiment derived from cited sources and external pages
Training Data (TR)
Pre-trained knowledge within the model
Generalized historical knowledge or learned associations
This distinction helps you understand:
Whether sentiment is influenced by current web content
Or by model training bias and long-term knowledge
Expanded brand view
Clicking a brand expands detailed sentiment insights.
Sentiment split by source
You’ll see sentiment separated between:
Web Grounding
Training Data
This shows:
Whether positive/negative perception is driven by live web content
Or embedded model knowledge
You’ll also see sentiment breakdown by AI engine.
Clicking an engine name filters charts and data for that specific model.
This helps you identify:
Engine-specific perception differences
Whether one model frames your brand more positively than others.
Sentiment over time
This chart shows the distribution of:
Positive keywords
Neutral keywords
Negative keywords
Over the selected timeframe.
Use this to detect:
Emerging negative narratives
Improving perception trends
Volatility across engines
Sentiment radar chart
The radar chart visualizes:
Most positive keywords
Most neutral keywords
Most negative keywords
This provides a quick visual summary of narrative drivers.
Keyword lists
Below the charts, you’ll see lists of all analyzed keywords.
Separated into:
Positive keywords
Neutral keywords
Negative keywords
For each keyword, you can see:
Number of times mentioned
Associated AI engine
This allows you to identify:
Recurring descriptive phrases
Reputation risks
Messaging alignment opportunities
Exporting sentiment data
You can export sentiment data to:
Google Sheets
CSV
Exports respect your selected filters and timeframe.
This is useful for:
Brand monitoring reports
PR analysis
Executive updates
External audits
How to interpret sentiment strategically
High visibility, negative sentiment
Your brand appears often, but language is critical.\
Focus on:
Addressing cited concerns
Improving messaging
Updating external content
Positive sentiment, low detection
You are described positively, but not often.
Focus on:
Increasing presence
Expanding prompt coverage
Improving citation footprint
Web Grounding negative, Training Data positive
Live web content may be influencing perception.
Review:
Recent articles
Reviews
News mentions
Training Data negative, Web Grounding neutral
Model bias may be influencing results.
Consider:
Increasing authoritative mentions
Improving long-term digital footprint
Relationship to other tabs
Overview page shows overall sentiment trends.
Competitors tab compares sentiment across brands.
Citations tab shows authority sources.
Search term detailed view shows sentiment in specific executions.
The Sentiment tab focuses specifically on narrative tone and descriptive context.