The AI Visibility page and its six tabs - Trackbase Academy

The dashboard gives you the headline numbers. The AI Visibility page in Trackbase is where you look behind them: which answers mention you, how they describe you, what the models searched for, which sources they cite, who you compete with and how you score per model.

The Mentions tab of the AI Visibility page in Trackbase, with total mentions, average position, mention rate and average sentiment above the mentions over time chart. Example data.

The six tabs

Tab	What it shows	Read more
Mentions	Every AI answer that mentions your brand, with total mentions, average position, mention rate and average sentiment. Filter by model, sentiment or answers with citations.	Mentions and citations
Sentiment	How AI answers frame your brand over time: the daily split between positive, neutral and negative, your current sentiment score out of 100, and the sentiment around citations.	Sentiment
Fan-outs	The searches AI models ran while answering your prompts.	Fan-out queries
Sources	The URLs and domains AI models cite, with the number of unique URLs, unique domains and the average number of sources per answer. Switch between a URL, domain and prompt view.	Mentions and citations
Competitors	The competitors you track, the ones Trackbase suggests and the ones you ignore.	Competitors
Heatmap	The visibility of your brand and every competitor, per AI model.	See below

Filters

Period: choose 7, 30 or 90 days, or set a custom range, at the top right.
Model: on the Mentions and Sources tabs you can limit the page to one model.

The heatmap

The Heatmap tab puts brands in rows and AI models in columns. Each cell shows a brand’s visibility on that model, and a darker cell means a higher score. Use it to answer two questions at once:

Where am I weak? Read your own row. A pale cell is a model that rarely mentions you.
Who is strong there? Read that column. The darkest cell shows which competitor that model prefers.

Switch between All and Tracked to include or leave out suggested competitors.

A good overall score can hide one model where you are nearly absent. Check the heatmap before you conclude that things are fine.

Last modified on November 12, 2026