Skip to main content
Overview is the screen a project opens on. It answers one question before any other: how often do the AI engines name you, in which of your questions, compared with whom, and did that change?

​
The toolbar

The toolbar states the scope every number on the screen is counted over, in one line: how many of your prompts are in your segments, how many answers were read, and over how many days. Below it, a line says when your history starts.
Which questions, segments, days and AI engines the numbers on this screen are counted over. Change the filters and every number here changes with them.
  • 7d · 30d · 90d picks the window. It starts on 30 days. The window you pick here is the one Brands, Prompts and Answers use too, so they always state the same days. Sources always shows the last 30 days.
  • all segments appears when you have claimed only some of your segments. It widens the scope to every segment. See Scope and segments.
  • Add prompt opens the sheet for adding a question of your own (see Prompts). An owner, admin or editor can add one.
  • Export is shown and not available yet.

​
What is on it, top to bottom

1

What moved

Only the changes big enough to be real rather than chance. If nothing is listed, nothing moved by more than its ± range. An empty list is a finding, and the screen says so rather than going blank. All activity opens Answers.
2

The metric rail

Your Visibility, Share of voice, Position when named and Sentiment when named, each with its ± range, the number of answers behind it (n) and its change since the previous period of the same length. Under the rail, your rank among every brand named in this scope.
3

Visibility over time

Your visibility and your competitors’, day by day, so you can see who is gaining ground and who is losing it. The chart draws up to four lines: you, the leader, and up to two brands just above you. A gap in a line is a day with no run, never a zero. Markers show the days the engines moved for every brand at once, and the days you published a change. In a project’s first days it says when the trend starts and shows where each brand stands on the first day instead.
4

Your visibility by model

Your visibility on each AI engine separately. Engines read different websites, so you can be strong on one and missing from another.
5

Setup

A few one-time steps that make these numbers more useful: add your own prompts, connect Google Analytics, invite a teammate, and turn on the daily alert email. None of them holds up collecting answers.
6

Segments

One row per segment: whether you have claimed it, how many prompts it holds, your visibility in it, its change, and the brand that leads it. A segment you have not claimed is shown, marked not in your scope, and not scored. Each change is against that segment’s own previous window.
7

Facets

Another way to slice the same questions: by the angle they are asked from rather than by topic. Each chip shows your visibility in it, its prompts and its leader. A muted chip has too few questions to read anything into.
8

Top losing prompts

The questions where you are furthest behind the brand named most. They are usually the best place to start. Each one opens its Prompt detail. If you are behind on none, the panel says so.
9

Alerts

A pointer to the Alerts screen, where fired alerts live. What moved, at the top of this page, is worked out from the figures here rather than from a rule.

​
How to read the numbers

  • Every figure carries its ± and its n. The ± is how far the true number could reasonably be from the one shown; n is how many answers it rests on. See The ± band and n.
  • A change smaller than its ± is shown muted, as “within noise”, with no arrow and no colour. It could be chance, so it never appears under What moved. See Within noise.
  • A dash with “needs one full run” means there are no answers to count yet. “0% · not named” means there were answers and none of them named you: that is a finding, not a gap in the data.
  • A low sample badge sits beside a segment’s visibility in the Segments table when it rests on only a few answers. The number is still shown; read it as an early signal.
  • Changes are in points. Going from one share to another is a change in percentage points, never a percentage of a percentage.

​
Before your numbers arrive

Overview has two early versions, so a new project is never a blank page.
No answers have been read yet. Overview shows what is being asked of which engines, each engine’s progress as a count, the questions being asked, and what will appear here once answers are read. The page checks again by itself. If collection has not started today, it says so.
Answers are in, and none of them names you. Overview leads with your visibility at 0% · not named and its n, then shows who the engines name instead, on each engine; How to get in, with steps built from your own data (a site you can get listed on, a question to win, and the Gaps queue); the pages the engines read, the ones you can get onto first; and a few of the latest answers with the brands they name marked.

​
Empty and error states

  • Nothing to read yet: the segments you have claimed have no prompts, or the window starts before collection began. The screen says which, and offers Show all segments.
  • Loading shows a placeholder in place of the blocks; the toolbar stays.
  • An error sits in the page with Try again. A block that fails on its own says so in place, and the rest of the page still loads.

Visibility

The headline number, and how it is counted.

Brands

The full league behind the rail.

Prompts

Every question, with its own figures.

Drift labels

What the markers on the chart mean.