Documentation
Visibility analytics and competitive intelligence
How to read the dashboard, mention trends, and competitor comparisons — from the big picture to precise growth zones
Open in dashboardHome dashboard: visibility KPIs in one view
The GEO Scout home page is a snapshot of your key visibility metrics in AI responses. Four KPI cards, a coverage trend chart, and a radar benchmark against competitors give an instant read: where you stand in AI now and where you're heading.
- Four KPI cards: Answer coverage — in how many prompts your brand appeared in the AI answer at all; Share of Voice — your share of mentions among all competitors; Domain citations — how often AI linked to your site; Citation Share — your domain's share of all cited domains
- Every KPI shows the change vs the previous period. The arrow and percentage tell you whether the metric is growing, falling, or flat. This is the main signal that your strategy is working right now
- The "Dynamics" block — a line chart of daily values with a metric switcher: answer coverage, Share of Voice, sentiment. Each metric shows its own angle on the trend — steady growth, zig-zag, decline, or stagnation — and the exact day the turn happened
- AI visibility benchmark — a radar chart pitting your brand against top competitors across five metrics at once. You instantly see where you lead and where competitors hold a clear advantage
Start your analysis with dynamics, not absolute values. A 13% SoV is excellent if it was 5% a month ago and alarming if it was 25%. Always look at the change indicator vs the previous period.
Competitive intelligence: market KPIs, trends, and gaps
The Overview tab of the Competitive intelligence page shows the whole competitive field: KPI cards, the dynamic split of Share of Voice between brands, mention trends for every competitor, and the list of prompts where you drop out of the AI answer.
- Four competitive-intelligence KPI cards: total competitors on the market, average mentions per competitor, your Share of Voice, and the total mention count of your brand — each showing the change vs the previous period
- Share of Voice dynamics — a stacked area chart of how every brand's SoV is distributed and shifts over time. You see at a glance who's eating your slice and whose share is shrinking in your favor
- Competitor mention trends — a line chart per brand. Hover any date for a tooltip with the exact mention count: map spikes and dips to specific events like releases, publications, or model updates
- Competitive gaps — prompts where AI mentions competitors but not your brand. For each gap you see which competitors made it in and across which prompts: precise growth zones, not vague "we're losing"
Don't try to close every gap at once. Sort by prompt frequency and intent priority (commercial "buy / order" beats informational), and start with the top three. Each closed gap is +N mentions to your Share of Voice.
Competitor detail: direct 1-on-1 comparison
Open a specific competitor's card for a detailed 1-on-1 comparison with your brand: positions, citations, sentiment, and mention trends. This is a decision-making screen — where exactly the competitor is ahead and on which metrics.
- Brand comparison — two columns with the key metrics: Share of Voice, mention count, average position in the AI answer, and domain citation count
- Mention sentiment — positive, neutral, and negative as percentages. Critical when SoVs are roughly equal: one brand is mentioned by AI in a positive tone while the other appears neutral or critical, and the customer picks the first
- A summary indicator — "Your brand is ahead by X%" or "behind by X%" — right under the comparison cards. A quick answer to whether you lead or trail against this competitor
- On the right, mention trends for the two brands only: yours vs the selected competitor. Spot where the competitor accelerated or slowed — leads for analyzing what actually worked for them
Sentiment is often underrated. With 20% SoV but 40% negative mentions, AI is mentioning you in a negative tone and the customer picks a competitor with lower SoV but a positive image. Working with negative sentiment is as much a part of GEO as growing the mention count.
Full competitor ranking
The "Top competitors" tab on the Competitive intelligence page holds the full ranking of every player AI mentions across your prompts. A detailed table for spotting specific targets: who to track, who to overtake, who's growing faster than you.
- Every key metric in one table: rank, trend, name, domain, mention count, Share of Voice, average answer position, recommendations, and domain citations
- Sort by any column. By mention count for the most visible players; by trend for the fastest climbers; by average position for who usually lands in the top lines; by domain citations for who AI references as a source most often
- Search by name or domain. Real rankings can hold hundreds of competitors (600+ in active niches); search helps you find a specific one — a new entrant from a report or a name from a case study
- Toggles on the right add or remove a competitor from your main comparisons. The enabled ones show up in dashboards, the radar benchmark, and KPIs; disable anyone who isn't a direct competitor (adjacent markets, irrelevant domains) to keep the metrics clean
Review the ranking in full every few weeks — AI keeps surfacing new competitors on your queries. Enable the ones truly in your market and strip out the noise. A cleaner list means sharper analytics and less noise in your KPIs.
Sources: where AI gets its facts
The Sources page shows the map of domains AI references when answering your prompts. Domains are grouped by class — competitors, catalogs, media, communities, expert, authoritative — and sorted by citation share. It's the answer to "where does AI get its facts and where do we need to land".
- Six source classes at the top — Competitors, Catalogs, Media, Communities, Expert, Authoritative. Each tile shows the number of domains and their share of all citations. You instantly see the niche structure: which type of source AI uses most often
- The domain table is sorted by citation share: at the top, the ones AI references most. Each row has a source-class tag (Competitors / Media / Communities / …), the number of unique URLs, the citation count, and the percentage of total volume
- Top filters — period (30 days by default), brand, and prompt group. Useful for narrowing to a specific cluster (e.g. "Finding web-development contractors") and seeing which domains are authoritative in that segment
- An arrow to the left of a domain expands the list of its unique URLs — you can see which exact pages are cited and on which prompts. It's the entry point for the work: where to land via publications, which catalogs to be listed in, which communities to monitor
These sources are exactly what the command center uses to build external publication recommendations: the domains AI already references become concrete actions — where to pitch a catalog listing, where to earn a media mention, which communities to anchor in.
Filters by AI provider and cluster
The top bar on every analytics page has two global filters: AI provider and prompt cluster. They work across every analytics section at once — dashboard, competitive intelligence, responses — and let you look at slices of data, not just the aggregate.
- AI provider filter: pick any subset of the 10 providers. Each provider shows your Share of Voice and coverage in it — instantly clear where you're strong and where you drop out of answers
- Prompt cluster filter: pick the business directions you need. Each cluster shows your SoV and coverage for that group. Metrics recompute for the selected subset
- Filters apply globally across every analytics section: once you pick providers or clusters, every KPI, chart, competitor ranking, and gap shows data for that subset only. Set it once — the slice applies everywhere
- Quick focus switch: one screen for the big picture, a couple of clicks for a narrow slice ("Perplexity only, commercial clusters only") without reloads and without losing context
Use filters as a diagnostic tool. An aggregate SoV may look average, but isolate a single cluster or provider and a drop or spike becomes obvious. It produces far sharper decisions than working off pool-wide averages.