See every time AI answersmention your brand.
Your buyers ask ChatGPT, not Google. The AI Citation Tracker runs the questions they actually type across seven AI surfaces — from a real user’s country, not an API — sorts every appearance into cited, mentioned, or awareness, scores your Brand Visibility 0–100, and tells you exactly what to adjust to win the citation back.
Watch it read every AI answer.
One prompt, seven surfaces, three kinds of visibility, one score. Pause or jump to any step to replay it.
One tracked prompt runs across seven AI surfaces — from a real user's country, not an API.
Seven surfaces. One scan.
A single tracked prompt runs against every answer engine your buyers use — the five conversational ones and both of Google’s new AI surfaces. Each is polled on its own, because a citation on Perplexity tells you nothing about what ChatGPT did.
Being seen comes in three kinds.
A link is only the loudest signal. An AI answer can carry your brand three different ways, and each one means something different for your pipeline. We separate all three so you never mistake a soft signal for a hard win.
Cited
The strongest signal: the engine put a clickable link to your page in its answer. We canonicalise the URL, attribute it to your domain (subdomains included), and map it back to the exact article that earned it.
Mentioned
The engine wrote your name into the answer without a link — the citation URL-matchers miss entirely. We read the answer text, match your brand by word boundary, and keep the ±60 characters around it as evidence.
Awareness
The engine named you from memory — it never ran a web search. That is baked-in brand awareness, the earliest signal there is. We flag it separately so a training-data mention never inflates your live citation rate.
One score. Three honest parts.
Brand Visibility is a single 0–100 line on your dashboard, tracked over time. But it’s never a black box — it opens into the three components that built it, so a dip always points at the lever that moved.
How often you're cited
The share of tracked prompts, across all seven surfaces, where an engine links your page. The hard, defensible number.
You vs everyone else
Of every brand named in the answers for your prompts, what fraction is you. This is where competitors quietly take ground.
Where in the answer
Being cited first reads very differently from being the fifth source. We weight the score by where you land in each answer.
It reads the answer, not just the links.
URL-matching sees domains. We read the sentence. That’s how the tracker surfaces the rivals you never told it about, the sources the engine actually trusts, and how your own published pages are performing.
Rivals you never named
We extract the actual company names written into each answer — not aggregator domains like G2 or Clutch — so real product rivals surface, then get proposed with a confidence score and a one-click "Track".
Who the engine trusts
Every cited URL is tagged Your brand · Competitor · Partner · Unaffiliated, so you can see whether an answer is dominated by rivals or by neutral third parties like Wikipedia.
Time-to-visibility per page
We tie citations back to the articles you publish: which pages get cited, how many days it took (TTV), and which published pages have earned zero citations so far.
A surface you own. A surface a rival took.
The same prompt — "best CRM for SaaS startups" — polled across the engines. The tracker keeps the full answer text as evidence, then pulls out exactly why one surface cited you and another reached for a competitor.
You own this answer
"For SaaS startups, Northwind is widely recommended for fast onboarding and clean data hygiene…"
- Hard citation: linked northwind.io at position 1
- Named in the first sentence, ahead of two rivals
- Cited page mapped back to your "CRM for SaaS" pillar
- Time-to-visibility: first cited 9 days after publish
- Source role of the other cited link: Unaffiliated (Wikipedia), not a competitor
A rival took this one
"The most recommended options are HubSpot, Salesforce and Pipedrive, with HubSpot favoured for early-stage teams."
- You aren’t cited or mentioned — HubSpot is cited at position 1
- Winning page: a HubSpot comparison table with startup-tier pricing
- Gap: you publish no head-to-head comparison for this query
- Auto-discovered rival "Folk" named here, awaiting your approval
- Citation Gap Brief queued: ship the comparison, then re-scan
The question your stack never answers.
Rank trackers and single-surface "AI Overview" tools each answer a slice. The seven questions below decide whether you’re winning the AI answer layer, and only a real citation tracker answers all of them.
It doesn’t replace your rank tracker. It answers the question your rank tracker can’t even see.
Every scan ends in a plan.
A thermometer tells you the temperature. A thermostat changes it. For any prompt a competitor owns, the tracker writes a Citation Gap Brief — and turns each gap into a specific, typed next step.
Scan
Your prompts run across seven surfaces from your buyer’s geo, every answer stored.
Score
Cited, mentioned and awareness roll into your Brand Visibility, with rivals surfaced from the prose.
Brief
For a prompt you lost, we read the winning page and write a typed action list to take it.
Ship
"Write a page" hands to the AEO writer. Re-scan next cycle and watch the score move.
Citation tracking, answered.
How is this different from Search Console or a rank tracker?
Search Console, GA, and Ahrefs are blind to AI answers. They can tell you a Google position, never whether ChatGPT linked you this morning. The AI Citation Tracker polls seven AI surfaces directly with the prompts your buyers actually type, extracts every citation and brand mention, and scores your visibility on surfaces the classic tools cannot see.
Which surfaces do you check?
Seven in one scan: ChatGPT, Perplexity, Gemini, Copilot, Grok, and both Google AI surfaces — AI Mode and AI Overviews. Most trackers watch a single surface (usually Google AI Overviews) and let you infer the rest. You can’t infer ChatGPT citation behaviour from an AI Overview; the prompts and the citation rules are different on each one.
Do you use the engines' official APIs?
No — and that’s deliberate. Each prompt is run through the real chat UI from a residential IP in the country and persona you set, so the answer is what an actual buyer in that market would see, not a sanitised API response. Every raw answer is stored verbatim with a timestamp as evidence.
What's the difference between cited, mentioned, and awareness?
Cited means a linked URL to your site. Mentioned means your name appears in the answer text with no link — a signal pure URL-matchers miss. Awareness means the engine named you from its training data without running a web search at all, the earliest sign your brand is becoming a default answer. We track all three separately so no soft signal inflates your hard citation rate.
How is the Brand Visibility score calculated?
It’s a 0–100 composite of three published components: citation rate (how often you’re cited across surfaces), share of voice (your slice of every brand named in the answers), and citation position (how high in the answer you land). You always see the three parts, not just the headline number, so you know exactly which lever to pull.
How does it find competitors I didn't list?
By reading the answer, not the link graph. A cheap language model pass extracts the real company names written into each response, filters out review platforms and aggregators, and surfaces the genuine product rivals named alongside you — each with a citation count, a confidence score, and a one-click button to start tracking it.
What do I actually do with a low score?
Every scan can hand you a Citation Gap Brief: an executive summary, the pages competitors are winning with, and a prioritised, typed action list — ship a comparison page, pitch a roundup, complete a review profile. It’s the thermostat, not the thermometer. And when the action is "write a page," it hands straight to the AEO writer. See /aeo.
Track your brand across every AI answer. Seven days free. Card required to lock in your founding price.