AI Visibility Tracker — Brand Mentions in Google AI Overview & Perplexity

Your buyers have stopped searching and started asking. When someone asks ChatGPT, Google AI Overview or Perplexity to recommend a tool like yours — does your name come up, or a competitor's?

This tells you, with an honest margin of error rather than a number that only looks certain.

No subscription. No API keys. Pay for the checks you run.

What you get

For every question, on every AI platform you pick:

• How often you are mentioned — as a rate across repeat checks, with a margin of error
• Whether your site is cited as a source, not just whether your name appears
• Which competitors come up instead, and how consistently
• What changed since last time — mentions up or down, new rivals, sources gained or lost
• What the gap is worth, using real monthly searches and what advertisers pay per click
• A shareable summary report — a standalone page with your overall mention rate, share of voice, the questions you are absent from ranked by value, and which assistants' citations are worth reading. Built from the measured numbers, not written by a language model.

The thing nobody else tells you

Ask an AI assistant the same question twice and you often get two different answers. Every other tracker asks once and reports the result as if it were a fact.

We ask several times and report what actually happened:

cited in 3 of 5 checks — 60%, somewhere between 23% and 88%

That range is the honest part. It is the difference between "we slipped this week" and "that is just normal variation", and you cannot tell them apart from a single check.

We also grade each platform separately. Google AI Overview and Perplexity are consistent enough to trust. ChatGPT will reliably tell you whether you get mentioned, but the sources it quotes change almost every time — so every result says which of its own numbers you should rely on, in plain English.

Try it free

Leave See a free sample first switched on and you get a complete example instantly — no charge, no setup. Turn it off when you want to measure your own brand.

Don't know what your buyers ask?

Most people guess badly at this. Leave the questions empty, enter one topic instead, and the questions people genuinely ask about it are found for you and ranked by buying intent.

Good for

• Answer Engine Optimisation (AEO) and Generative Engine Optimisation (GEO) — knowing where you stand before you spend on fixing it
• Proving to a client or a board that AI visibility work is or is not landing
• Catching a competitor the moment they start appearing in answers you used to own
• Deciding which topics deserve content next, weighted by what they are actually worth
• Watching a launch land — schedule it weekly and the trend builds itself

Example

Input

{
  "brand": "Notion",
  "brandDomain": "notion.so",
  "seedTopic": "project management software",
  "samples": 5,
  "surfaces": ["aiOverview", "perplexity"],
  "weightByDemand": true
}

Output — one row per question, per platform

{
  "brand": "Notion",
  "query": "What is the best project management software?",
  "surface": "perplexity",
  "samples": 5,
  "mentionedIn": 3,
  "mentionRate": 0.6,
  "ciLow": 0.23,
  "ciHigh": 0.88,
  "confidence": "stable",
  "citationReliability": "reliable",
  "domainCitedIn": 2,
  "competitors": ["asana (5/5)", "monday.com (5/5)", "clickup (3/5)"],
  "citedDomains": ["reddit.com", "asana.com", "monday.com"],
  "previousMentionRate": 0.8,
  "mentionRateChange": -0.2,
  "newCompetitors": ["linear"],
  "monthlySearches": 33100,
  "avgCpcUsd": 22.4,
  "missedSearchesPerMonth": 13240,
  "gapValueUsd": 59315.2
}

Export to JSON, CSV or Excel, or feed it into a sheet, a dashboard or your own reporting.

Frequently asked questions

Is the report AI-generated? No. It is computed directly from the measured results — no language model touches it, so it cannot invent a number. A model is used only to read brand names out of each AI answer.

How is this different from the other AI visibility tools? Two things. Everyone else asks once and reports a rank; we ask several times and give you a rate with a margin of error, because the underlying answers genuinely move. And we attach real search volume and cost-per-click, so a gap comes with a number attached instead of a score out of 100.

Which AI platforms are covered? Google AI Overview, Perplexity, ChatGPT, Google Gemini, Microsoft Copilot and Google AI Mode. The first two are on by default because they are the most consistent.

Do I need an OpenAI key or any other account? No. Nothing to sign up for, nothing to configure, no keys of your own.

Why does ChatGPT show as less reliable? Because it is, and we would rather say so. Ask it the same question repeatedly and the websites it quotes are almost entirely different each time, though whether it names your brand is far steadier. So mention rates from ChatGPT are usable; its source list is not, and each result says which is which.

How many checks should I run per question? Three is a good default. One gives you an impression with no error bar. Five narrows the range noticeably. Ten is rarely worth the wait.

Can I track competitors as well? Yes, and you do not have to name them — rivals are picked up from whatever the answers actually recommend, which regularly surfaces names people were not expecting.

Can I run it automatically? Yes. Schedule it weekly and each run reports what moved since the last one. Your history stays on your own account.

Does it work in other languages and countries? Yes. Answers differ by market, so set the country your customers are in. English gives the richest AI answers today.

Also worth a look

Google Rank Tracker — the classic other half: where you rank in ordinary Google results, with the same real search volume and cost-per-click attached.