How we measure
We sell analysis, so the method should be inspectable. This page describes exactly what a check does, what the numbers mean, and where our confidence ends.
Last reviewed: 8 September 2026
Which AI engines we query
- Perplexity Sonar: used on every check, including anonymous ones. It searches the live web and returns the sources it used.
- OpenAI web search: added when you are signed in, as a second opinion from a different vendor.
- Gemini with Google Search: added on paid plans, a third opinion grounded in Google's own search index.
The consumer ChatGPT app and Google AI Overviews are not measured yet. Those products do not expose a way to measure what an individual user sees, and none of our engines reproduces them: "Gemini with Google Search" is an API answer grounded in Google Search results, not the AI Overview a signed-in Google user sees. If a reliable way to measure those surfaces appears, we will add it and say so here.
How a visibility check runs
- We fetch your homepage and sitemap and infer your category, offerings and market.
- We draft buyer questions that never mention your brand: 2 on an anonymous check, 5 when signed in. You can edit or replace every one before the run.
- You set the country, and optionally a city, suburb or postcode, so the engine answers for the area you actually sell into.
- Each question is asked once per engine. We record whether your brand was named, in what position, which competitors were named, and every source the engine cited.
- When a competitor is cited and you are not, we fetch their cited page and your closest equivalent and compare the signals below.
Because each question is asked once, a result is a live snapshot, not an average. AI answers vary between runs. Treat a single absence as a signal to investigate, and a repeated absence across runs as a finding.
What the audit scores
A site audit crawls the pages you select and scores seven signal groups. Each group is scored from the evidence found on the pages, then weighted into one readiness score. The report shows the arithmetic for every deduction, page by page, so you can check our work.
Organization and page-level JSON-LD: is there a machine-readable statement of who you are, what you sell and where, and is it valid?
Whether your name, description and contact details say the same thing across your meta tags, Open Graph tags, JSON-LD, footer and About page.
Depth, numeric specificity and whether a page answers a real buyer question rather than restating a category.
Articles are scored on bylines, dates and outbound citations. Company pages are scored instead on organisation schema depth, named leadership, proof, reviews, reachable contact details and policy pages.
Titles, descriptions, canonicals, indexability, image alt coverage, and whether your pages need JavaScript to show their content.
How similar your pages are to each other once shared headers and footers are stripped out. Shared navigation is not counted as duplicate content.
Whether the identity in your schema matches the identity in your visible copy.
What we do not claim
- We do not claim these signals are confirmed ranking rules. No AI vendor publishes how its retrieval picks sources. Our scoring reflects patterns we observe across audits and the public documentation that does exist.
- We cannot promise placement in an AI answer, and nobody can. These are synthesized results, not ad slots.
- A high readiness score means your site is well set up to be cited. It is not a prediction that you will be.
- We do not measure the consumer ChatGPT app or Google AI Overviews, and we do not estimate them from the engines we do measure.
What a free account includes
2 buyer questions per anonymous check, 3 credits a month once you sign in (one full 5-question run). Anonymous checks do not need an account. See pricing for what paid plans add.