A dashboard is a name, a window, and the tiles you chose.
Reporting on AI visibility usually means somebody rebuilding the same slide every Monday. A saved view replaces that: build it once, and the exec, the client and the person who joined last week are all reading the same thing, scoped the same way.
A dashboard stores configuration, never a number
Every tile reads the same endpoint as the screen that metric already lives on. That is a deliberate boundary: there is no second metrics pipeline behind dashboards, so a tile can never quietly disagree with the page it came from.
The window is saved on the view, not in the link
A saved view whose date range lives in the URL is not really saved. The range and the comparison belong to the dashboard, so the person you send it to sees the window you built, not their own leftover filter.
Filters are only the ones that really apply
Segment, engines and competitors, because those are the dimensions the underlying endpoints can honor. Prompt tags are deliberately not offered: a chip that looks like it filters while changing nothing would leave you reading numbers you believe are scoped and are not.
Zero and not measured are different
A tile with no runs behind it says so rather than drawing a confident 0.0%. An empty state is information; a false zero is a wrong answer.
Ten types, every one backed by a metric you already have.
Each tile renders as a chart or a table where both make sense, and carries its own filters, so one dashboard can hold a company-wide view and a segment view side by side.
Visibility Score
Your standing across the prompts you track, over time.
Share of Voice
Your slice of the answers against the competitors you name.
Citation Share
How often the pages you own are the ones an engine cites.
Average Position
Where you tend to land when an answer ranks its picks.
Prompt and Engine Matrix
Every tracked prompt against every engine, shaded by the metric you pick.
Ranking Leaderboard
Where you place against competitors across the prompts you track.
Top Cited Sources
The pages and publications the models lean on in your category.
Sentiment Trend
How the tone of what models say about you moves over the window.
Needs Perception turned on
Sentiment Score
The current reading, per engine.
Needs Perception turned on
AI Crawler Activity
How often AI crawlers fetched your pages, and how many fetches failed.
Needs a connected log source
Saved for the workspace, credited to whoever built it.
Dashboards live with your organization rather than in one person's browser. Duplicate one to fork a variant, rename it, and the original stays where the rest of the team left it. Nothing on a dashboard escapes your workspace: the tiles read your data through the same permissions as every other screen.