Overview
Overview answers three questions about one project on a single page: how big the AI market you monitor is and what you get from it, what AI says about your brand, and how you compare with the competition. It is a reading page. Nothing on it is calculated in a new way: every figure comes from Visibility, AI Ads, Brand analysis, Traffic (Google Analytics 4) or the WWW audit, so a number here always matches the number in its module. If it does not, that is a defect worth reporting, not a different point of view.
Open Overview at the top of the Monitoring group in the sidebar.
What the page reads
Section titled “What the page reads”Open a project, then choose the model and the period at the top of the page. Overview reads one model at a time or all models together. The period is remembered separately for this page, so a period picked here does not change what Visibility or AI Ads show. The model is shared with Visibility: picking one model here picks it there too. When Visibility is set to several models, Overview reads all models together. The comparison baseline (previous period) drives every “vs previous period” note on the page.
The line under the title is a status bar: your position among the brands in your category, the distance to the leader, the biggest opportunity found in the competition section, and, in red, the number of answers in which AI advises against your brand. Each member appears only when its data exists, so a project without a brand analysis run shows a shorter line, not a zero.
How to read a number
Section titled “How to read a number”Every figure has three helpers next to it:
- A tag on each Market tile says what kind of number it is: estimate for the AI query market (derived from the monthly query volume attributed to your prompts), measurement for visibility (counted from the answers PromptEye checked), measurement + estimate for visits and GA4 measurement for leads.
- The info icon next to a card title explains how the figure is counted, which module and data it comes from, and which period it covers. Everyone reading the page sees the same definition.
- The change is compared with the previous period or, for brand analysis, with the previous run. Visibility and visits are coloured only when the change is large enough to be meaningful for that metric; a small change stays grey with its number still visible. A change is never hidden to make the page look better.
Units are always percentages. “34% visibility” means 34% of the checked AI answers mentioned your brand; a change of “+2.1%” is a difference of percentage points on that same scale.
Market
Section titled “Market”The first row shows how the monitored AI market turns into results: the estimated number of AI queries your prompts represent per month, the share of answers that mention your brand, visits from AI assistants and leads from those visits. With fewer than five lead events the leads tile says there are too few to judge instead of showing a change. Visits and leads come from a connected Google Analytics 4 property. When the project has no GA4 connection, those two tiles are not shown at all; connect GA4 in Integrations to see them.
The visibility tile also tells you how visible the leading competitor is (or that you lead, and who follows), and, when your own ads were observed in the period, in what share of answers your ad appeared. The number of answers with your brand is derived from the visibility share. Ads are never added to visibility: an ad next to an answer is not a mention. When the ad figures could be read from only part of the period, a note under them says so.
From AI queries to leads
Section titled “From AI queries to leads”Under the tiles, a flow chart follows the same numbers from left to right: the AI query market for the period, the answers with your brand, with your own ad and without your brand, then visits to your site and the answers seen without a click, and finally leads and visits without a lead. Each ribbon carries part of a box to the next step; the boxes have a fixed height, so read the numbers, not the area. Answers without your brand fade out to the right: that is the room to grow. The ad box ends there too, because the chart does not follow clicks on ads separately.
- Groups. The arrow on the market box splits it into your prompt groups (the six largest and the rest together), so you can see which group feeds the answers with your brand.
- Hover a box or a ribbon to highlight its path through the chart and read the exact number.
- Click a box to open its details with the exact figure and a link to the module that explains it: Prompt Intelligence for the market and groups, Visibility for answers, AI Traffic for visits and leads.
- The timeline under the chart shows daily site traffic from AI and leads over the last fifteen months. Drag the highlighted window or its edges to choose a period of 7 to 60 days; the whole page switches to that custom period. Picking a period at the top of the page moves the window.
Without Google Analytics 4 the chart stops at the answers and the timeline is hidden. Without a query estimate for your prompts the chart is not shown.
Brand — what AI says about us
Section titled “Brand — what AI says about us”This section reads the newest ready Brand analysis run and the sources cited in the selected period.
- How does AI speak about us? classifies the tone of answers that mention your brand: positive, rather positive, neutral, mixed or negative, with the share of positive answers shown as a bar and the full split listed under it. The verdict needs a minimum number of answers with your brand; with fewer, the card says there are too few answers to judge instead of guessing.
- Does it recommend us? shows how often the brand is recommended, compared or only mentioned. When some answers advise against your brand, a warning names how many; advising against is a separate signal, not one of the roles. This is the number the red member of the status bar refers to.
- What does it praise, what does it blame? lists the leading positive and negative attributes named in the same run. An empty column says there was nothing to list in this run.
- Where does AI know us from? ranks the domains cited in answers about your brand. Your own domains are marked, and the summary tells you what share of citations point to your pages. If no source was cited in the period, the card shows the verdict only.
If there is no brand analysis run yet, the tone and role cards show their “no analysis” verdicts and nothing else. Start a run in Brand analysis to fill them.
Competition
Section titled “Competition”The section sentence tells you your position, the distance to the brand directly above you and to the leader, the share of mentions held by the three leaders, and how often their pages are cited compared with yours. When AI does not mention your brand in the monitored questions at all, the sentence says so and the table shows the leaders only.
- Your competition shows the top three brands and the brands directly around your position, marked You. A line between them counts the brands that were skipped. A brand marked “outside the list” appears in answers although you did not add it as a competitor. An Ads tag marks brands whose ads were observed in the period.
- Share of mentions splits the mentions in the period between the leading brands, you and the rest.
- Share of ads does the same for observed ad impressions and adds whether anybody advertises on your brand name. When nobody advertised in the category in the period, the card shows a single line saying so; a category without ads is information, not missing data.
- Why they rank higher compares you with the three leaders on three causes: whether their pages are cited as sources, whether their ads sit next to decision prompts, and whether the newest WWW audit found technical flags against you. Each card says whether there is a gap, no difference or an advantage on your side, and the header line names the biggest gap and the prompt group where it is widest. The ads card counts decision prompts where a leader advertises and you do not; the technique card counts the checks the newest audit flagged against you. A cause without data for this project says it is not measured yet.
Common questions
Section titled “Common questions”Why does the page show “We could not load this section”? The figures for the chosen period could not be calculated. Use Try again. If Visibility is being recalculated after a change to the brand definition or competitor groups, try again once the rebuild finishes.
Why does the page show “We could not read visibility for this period”? The figures that come from AI answers (visibility, the competitor ranking, sources) could not be read for the chosen period. Traffic, leads and ads still load, and the cards that need visibility show no data until you use Try again.
Why is there a “Data through” note under the header? Visibility figures include today’s answers as of the last hourly refresh. When a refresh is late, the note says which day the figures reach; later days appear after the next refresh.
Why is a number different from the one in its module? It should never be. Check that both pages use the same model and period first; if they do and the numbers still differ, report it.
Why do I see fewer tiles or cards than described here? Tiles and cards depend on connected data: Google Analytics 4 for visits and leads, a Brand analysis run for tone, role and words, observed ads for the share of ads, a WWW audit for the technical cause. Missing data hides the element instead of showing a placeholder value.
Can I click into the details? In the flow chart, yes: a click on a box opens its details and a link to the module behind it. The other cards show evidence only; open the module named under the info icon to investigate an answer, a source or a competitor.