The Sentiment Audit shows per-Topic brand sentiment across the Personas in your Project and the AI models they ask. It's included in the Playbook (also known as the Pro plan), and you can add it to Starter as your one added audit with Build your own. See What's included in the Starter, Build your own, and Playbook plans?
Where do I find the Sentiment Audit?
Open your Project and select Sentiment Audit in the left menu, under Project.
How does the Sentiment Audit run?
The Sentiment Audit runs automatically whenever your Project runs, on your subscription schedule. It can't be rerun on its own. When a run starts, it typically takes from about 30 seconds to a few minutes, and the page updates live as results come in.
If you want fresh results sooner, change your Project's run day to tomorrow in Settings. See How often does my Visibility Audit run, and can I change the schedule? Or ask for a human in chat: our support team can run your whole Project for you at the usual cost.
What views are available?
- Themes: recurring themes in how the models talk about the selected brand, grouped into Pros, Cons, and Caveats, each with a count. Click a theme to see the questions behind it.
- Topics: a bubble chart placing each Topic by sentiment (vertical) against visibility (horizontal), with bubble size showing the number of sentiment questions behind it. Click a bubble to see the questions behind it.
- Persona × Topic, Model × Topic, and Stage × Topic: heatmap-style tables showing sentiment on a -1 to 1 scale (red is negative, green is positive), averaged across the dimension not shown, with each cell noting how many observations it's based on. Click a cell, row, or column header to filter the questions below to that slice.
How do I see the underlying questions and answers?
The Sentiment Audit has its own Conversations page, reached from Conversations under Sentiment Audit in the left menu. This is separate from your Visibility Audit's Conversations page: each one shows only the questions and answers behind its own audit.
On this page, you can filter by Topic, Persona, Model, Stage, and Theme, and a running count at the top shows how many questions and answers currently match your filters.
Above the individual questions, a Topic breakdown shows the sentiment score for each Topic across every model included in your Project, plus an overall row.
Below that, each individual question is listed with the Persona who asked it, its Topic, and its buyer-journey Stage (for example Product / Vendor Aware or Evaluation / Decision), along with every model's sentiment score for that question. Click any model's score to open its full response in a panel next to the question.
What should I do next?
Start with the Cons and Caveats in Themes, since they show why AI models hesitate to recommend you. Then fix what they point to with content: the Content Audit shows which pages to create or improve. See What is a Content Audit?