How MentionScout works

MentionScout tracks how your brand shows up when people ask AI answer engines questions. You tell it what questions to watch, it runs those questions against the major models on a schedule, and it records every time your brand is named or your pages are cited. This page explains the model so the rest of the docs make sense.

​

The core model

Four things connect in a chain. Everything in the app is a view on one of them. So the flow is: you track prompts -> MentionScout runs them on a schedule -> each run produces mentions and citations. Your visibility score, sentiment, and competitor comparisons are all aggregates over those runs.

​

Prompts run on a schedule

Each prompt has a frequency of daily, weekly, or monthly, and MentionScout schedules the next run accordingly. You can also run any prompt on demand and see results within a minute or two. Prompts have a type that shapes how they are read:

​

Mentions vs citations

This is the distinction the whole product hinges on, so it is worth being precise. A single run can produce any combination of the two:

• Mention, no citation: the answer says good things about you but links someone else’s page as the source.
• Citation, no mention: the answer links one of your pages but never names your brand in the prose.
• Both: the answer names you and cites your page. This is the strongest result.
• Neither: you did not show up in that answer at all.

Tracking them separately is what lets you see, for example, that AI recommends you (mentions) but sends readers to a competitor’s page for the details (citations). Mentions and citations each have their own page in AI Visibility.

​

How the app is organized

The sidebar groups features into three ideas. Knowing the mental model makes the navigation obvious.

​

The Visibility overview

The dashboard rolls all of this up for the brand and time window you pick. Use the range control at the top right to switch between the last 7, 30, or 90 days. The three headline numbers at the top are:

• Visibility score: the share of completed runs that mentioned your brand, as a percentage.
• Sentiment score: how positive the model was when it did mention you, on a 0 to 100 scale. It reads n/a until you have at least one mention.
• Runs: how many runs completed in the window.

Below those, the dashboard breaks the picture down:

• Brand visibility charts your share of runs against each competitor you track.
• Sentiment when mentioned splits sentiment by prompt type and trends it over time.
• Competitor visibility heatmap shows visibility percent for you and each competitor on every model, side by side, so you can spot the engines where a rival outranks you.
• Visibility over time trends your daily share of mentioning runs.
• Top cited domains and Top ranked URLs list the sources models lean on, and which of your own pages get cited.
• Recent mentions and Recent social mentions surface the latest hits from AI answers and from the community.

​

The crawl-to-click funnel

Mentions and citations tell you what AI says. The funnel tells you what happens around it: AI crawlers fetch your pages, models cite some of them in answers, and readers click through to your site. The Crawler activity card on the dashboard summarizes the last 30 days as two numbers:

• AI crawls: how often AI assistants fetched your pages.
• AI-referral clicks: how many visits those AI answers sent back to you.

Together with citations, this closes the loop from “AI read my page” to “AI cited my page” to “a person clicked through.” Connect it once and it runs on its own.