The Overview dashboard

The Overview is the home page of GEOflux: everything collected from your prompts, counted and charted. It is split into tabs, each answering one question, and the filters in the top bar scope every tab at once. Set the scope first, then read: a number only means something for the period, engines and prompts it was computed over. See Filters and time periods.

Before your first answers

A new project has no answers yet, so the Overview shows a first-run screen instead of the tabs: your suggested prompts and competitors while they are being prepared, then Your first answers are on their way while the engines answer. The first answers usually arrive within about 10 minutes of tracking a prompt, and the tabs appear on their own. See the Quickstart.

How to read any tab

The lead strip answers the tab's question first. Most tabs open with a strip of headline cards before the detail. With a single brand selected they describe that brand; with All brands selected they describe the market (see The All brands view).
The brand in focus is orange. That is your own brand unless you pick another one in the Brand filter. When you focus a competitor, a banner says so and offers a switch back. Every other brand keeps a stable color taken from its logo, so you can follow a series from chart to chart.
Click a brand to focus it. Clicking a brand in a chart or a table sets the Brand filter to it.
Percentages have one decimal. The change since the previous period of the same length is shown as a signed value, for example +2.3%: that is the difference in percentage points, so 10% to 12.3% reads +2.3%. A gap between two brands is labelled in points (pp). With nothing earlier to compare with, the change reads no baseline; on Free, it reads too few runs until the period holds at least 3 runs.
Zero is not the same as no data. A real zero is a finding: you were eligible and not named. An empty chart means nothing was collected for that slice.
Hover for the exact value and date. Points can carry chart notes explaining a spike or a drop.
Today can be incomplete. While today's run is still arriving, a notice above over-time charts says so. See When data updates.

The tabs

Visibility

"Are we showing up?" The default tab.

Visibility over time: the share of AI answers that name each brand, Over time or By brand. See Visibility.
Mentions: how many answers name each brand, By brand or Over time. See Mentions.
Prompt coverage: the share of your prompts on which each brand is named at least once. See Prompt coverage.
Rankings: every tracked brand with Visibility, Share of Voice, Sentiment and Mentions. Click a header to sort. The Share of Voice column here is always measured against all brands.

"When AI names brands in our space, how often is it us?" A Composition donut or an Over time chart, then Rankings sorted by share of voice. The toggle in the lead strip sets the comparison: vs. competitors (the default) or vs. all brands. See Share of voice.

Sentiment

"When AI names us, how does it talk about us?" Scores run from 0 to 10: below 4 is negative, 4 up to 7 is neutral, 7 and above is positive. See Sentiment.

Sentiment Score: every brand side by side (Per-brand bars) or your score over time (Sentiment trend), with the zones shaded behind.
What's driving this: the attributes pulling your sentiment down and holding it up, for example "Pricing & value" or "Customer support". Part of Brand Perception.
Sentiment by topic: which topics read well or poorly, as a treemap or a list, once your prompts have topics.
Brand Perception: what AI says you are good and bad at, attribute by attribute, with the answers behind each one and a benchmark against your competitors. See Brand perception.

Brand Perception is included on Starter, Growth and trials. On Free, both parts show what they would add, with a lock.

Sources

"Which websites does AI rely on, and is our site one of them?"

The lead shows Your content cited (citations of your own pages and your citation rate), the Leading source, and sources New this period.
Where Competitors Are Cited and You Aren't: sites whose answers name a competitor but never you, grouped by type of site. Your shortlist for PR and outreach.
Sources per mention: how much evidence AI attaches to each brand when it names it.
Top Domains: the most cited domains, as a trend or a list.
Citations by Domain Type: what kind of sites are cited and whose they are (yours, competitors', third-party), as a chart, a trend or a list.
All Domains / All URLs: the full table, with Type, Used, Citations and First seen, filterable by site type. It shows up to the top 1,000 sources by citations.

Click any domain or page for its detail page. See Understanding sources and Citations.

Brands

"Who are we really competing with in AI answers?" Every brand the answers name, tracked or not; an untracked brand is listed once it appears in at least two answers in the period.

The lead shows your share of voice, how many brands were mentioned, and how many untracked brands are surfacing.
Top Brands: the most mentioned brands, as a trend or a list.
Appears alongside: the brands AI most often names in the same answers as you.
All Brands: the full table, with Mentions, Share of voice, Sentiment and First seen, and You and Tracked badges. A row's menu opens the brand's profile, or tracks or untracks it as a competitor.

See Track competitors.

Personas

"Does AI answer differently for different buyers?" Visibility, sentiment and share of voice per persona, each against Base (the answer with no persona). The lead names the persona you are most and least visible to; a card per persona links to its answers, and a comparison chart sets the brands side by side across personas. The tab needs a single brand, so under All brands it asks you to pick one. On Free, it shows a preview. See Personas.

Topics

"Which themes do we win, and which do we lose?" Built from the topic on each prompt, so it asks you to add topics until you have some.

An insight strip opens the tab.
Performance by Engine: a heatmap of visibility per topic and engine, colored by sentiment. A flag marks topics where the engines disagree most.
Topic Distribution: a treemap where tile size is the number of prompts and color is sentiment.
Competitor breakdown by topic: how each tracked brand does on every topic, with its own share of voice toggle. It needs a single brand.
Topic Leaderboard: every topic ranked. Expand one to see its prompts.

See Organize your prompts.

Buyer Journey

"Where along the way do we drop out?" Built from the Buyer Journey step on each prompt: Discover, Evaluate and Buy.

Buyer Journey: a funnel with one band per step, for Visibility, Share of voice or Sentiment. Each step shows its prompt and answer counts, your value with its change against the previous period, the change from the step before, and a daily trend with the step's leader as a dashed line. Under All brands it shows each step's leader instead.
Rankings by step: the brands ranked within each step.

The Buyer Journey filter chooses which steps appear. Share of voice on this tab is always measured against all brands. Prompts without a step are left out: if the tab reads No Buyer Journey steps yet, use Assign Buyer Journey steps to set them on the Prompts page. See Organize your prompts.

Engines

"Which engines name us, and where do we trail?"

The lead summarizes Across engines.
Which AI engines mention you most?: a card per engine with your visibility, mentions and sentiment there, marked Strong or Needs attention. An engine left out by the Engine filter reads Not selected.
Engine trends: each engine over time, for Visibility, Mentions or Sentiment.
Competitor breakdown: you against each competitor, engine by engine, with a share of voice toggle.

Use it to find the one engine where a competitor beats you, then read that engine's answers on the Answers page.

The Share this view button at the top of the Overview turns what you are looking at into a shared report: a read-only page at a link, for someone without a login. The report carries the period and the Engine, Topic, Tag and Persona filters. A preset period keeps rolling; a custom date range stays fixed. A report is always about your own brand, so the Brand filter is not carried.

Share is disabled, with a tooltip, on the Buyer Journey tab and while a Buyer Journey or Branded vs Non-branded filter is set, because reports cannot carry those yet.

Good practice

Fix one headline scope and keep it, for example last 30 days, all engines, Branded vs Non-branded set to Non-branded. Read it the same way every time so numbers stay comparable.
Read in this order: Visibility (are we there), Share of Voice (against whom), Sentiment and What's driving this (how we are described), Sources (why), then Topics and Buyer Journey (where to act).
Change one filter at a time when a number surprises you, and watch which one moves it.
Drill into evidence before acting. Every chart leads to answers: open the Answers page, a prompt, a brand or a source to see what is behind the number.

The Overview dashboard

On this page