Can a free tool measure my AI visibility?

No. A free scan can tell you whether an AI engine can read your page and find something worth quoting. This is real, useful information, and it’s why we offer one. What no scan can tell you is whether engines actually name you when your customers ask. Visibility lives inside the engine’s answers, and the only way to measure it is to ask.

What a page scan can see

A scan fetches one page the way a crawler does and checks what’s checkable. Can the content be read without running scripts? Are the headings real? Is there a clear, quotable answer on the page? Does the structured data match the visible copy? These checks are honest work. Run them twice and you get the same result twice. They cover the first two of the three things an engine needs from you: readability and citability.

What no crawl can see

Visibility is different. Whether an engine names you depends on things your page doesn’t hold. What the engine already knows about you. What other sources say. Who else is competing for the same answer. A crawl of your own page can’t see any of this. It’s like judging a restaurant’s popularity by inspecting its kitchen. The kitchen matters. It just isn’t attendance.

The proof: 30 to 1

In January 2026, researcher Amit Prakash Sharma published a study with a finding every business should pay attention to. He took 112 startups and ran 2,240 questions through ChatGPT and Perplexity. When the engines were asked about the startups by name, they recognized them almost every time: 99.4 percent for ChatGPT. When the engines got discovery questions instead, the kind your customers actually ask, like “what are the best tools for this job?”, the same startups showed up just 3.32 percent of the time.

The engines knew these companies. They just never brought them up. Being known and being named are different things, and in these data, the gap was 30 to 1.

The score that predicted nothing

The same study checked one more thing. Did optimization scores, the kind that page-scanning tools produce, predict who got recommended? They didn’t. No correlation at all. What did predict discovery: links from other sites and presence in real communities. In plain words, outside evidence. Engines recommend who the wider web backs up, which is the same reason they recommend your competitors when the outside evidence favors them. A score computed from your own page can’t contain any of this. It doesn’t matter how carefully the score is computed. The inputs it needs aren’t on the page.

A two-minute reality check

You can run this study’s core test on your own business right now. Open an engine and ask about your company by name. It probably knows you. Now ask the question a customer would ask, in their words, with no brand name in it. See if you come up. You just measured your own discovery gap. And notice what happened if the engine knew you but didn’t name you: your pages might scan perfectly, but no page score anywhere would have warned you.

What ours claims, and what it doesn’t

We offer one of these free tools ourselves, so here’s the honest label. Our scan is a single-page scan for readability and citability. It asks whether an engine can read you and quote you, and it answers this question well. It does not output a visibility score, because a page crawl can’t measure visibility, and we’d rather be right than impressive. When any tool hands you an “AI visibility score” after crawling one page, it’s measuring the kitchen and calling it attendance.

How to actually measure visibility

Ask engines. That’s the whole method. Take five questions your customers ask, in their words, with no brand names in them. Ask each one across ChatGPT, Claude, Perplexity, Gemini, and Grok. Write down who gets named, who gets cited, and who’s absent. Then repeat it next month, with the same questions. A single prompt tells you very little. A repeatable benchmark across engines, questions, and time tells you where you stand and which way you’re moving.

What to do about it

Use both measurements, in order. Run the free scan first, because if a crawler can’t read your page, nothing downstream matters. Then measure the answers themselves: by hand with the five-question method above, or properly with an AI Visibility Audit, which runs the questions your customers actually ask across engines and benchmarks you against whoever is winning the answer today.

A scan can tell you whether you’re quotable. Only the engines can tell you whether you’re named. Never let one number stand in for the other.