Skip to main content
Every scan collects the full text of each AI response to your tracked prompts. The Mentions section is where you read the source material: exactly what AI Overview, AI Mode, ChatGPT and Gemini said, with your brand and your competitors highlighted in context. When a score moves in Analytics, this is where you come to find out why.

​
The Mentions view

Open Mentions → All Mentions to see every response Mentioned has collected for the selected date range. The date-range selector at the top-right of the dashboard applies here, so you can focus on any period you like. At the top of the view, four summary cards show your totals: Total, Positive, Neutral and Negative mentions. Click a card to filter the table below to just those mentions.

​
The results table

Each row in the table is one tracked prompt from one scan, with a column for each AI engine. The engine cells tell you the outcome at a glance:
  • A green dot means your brand was mentioned in that engine’s response; a grey dot means it wasn’t.
  • Where the response ranked brands, the position your brand held is shown alongside the dot.

​
Reading a full response

Click any row to open the response panel. This shows the actual text each AI engine returned, not a summary, with inline highlighting so the important parts stand out:
  • Your brand is highlighted in green.
  • Competitor brands are highlighted in amber.
  • A sentiment badge shows whether the mention was positive, neutral or negative.
  • Cited sources lists the web pages the AI referenced in its answer, so you can see which content is shaping the response.
Because every engine’s response to the same prompt sits side by side, you can compare how each one treats your brand.

​
How sentiment is judged

Sentiment is assessed by Mentioned’s analysis of the full response, not just the sentence containing your brand name. A mention is positive when the response frames your brand favourably (recommending your product, citing your expertise), neutral when it’s factual or descriptive, and negative when it carries critical language or comparisons that disadvantage your brand.

​
Tracked queries

Mentions → Tracked focuses the same data on your tracked prompt list, so you can review performance prompt by prompt: which questions consistently surface your brand, and which ones the AI engines answer without you.
When a prompt never returns your brand, read the full responses for it before changing anything. Often the AI is recommending competitors based on content the engines have seen, which tells you exactly what kind of content you need to create to compete for that answer.
If your results look unexpected, check the prompt phrasing first. Small wording changes can produce very different AI answers. See How scans work for what Mentioned records from each response.