AI Visibility Checker

What this score is, and what it is not

This score is a sum of the eight answers you just gave. It is your own read of your own business, added up honestly. It is a readiness checklist, not a prediction. It does not tell you where you will rank inside an AI answer, how many times you will be mentioned, how much traffic any of it will send you, or what it is worth in revenue. Nothing here measures a real outcome, because the tool has no access to your analytics, your mentions, or the answers an assistant actually gives about you. Two businesses with different results can answer these eight questions identically and land on the same number.

What it does tell you is whether the raw material an assistant needs is present and consistent. That is a real property of your business, it is not a guess, and it is worth fixing whether or not you ever see a number move.

How to read the eight questions, and why each one is in the list

Question 1, the one sentence description, is the highest weight at 14 points because it is the raw material for almost everything else. If a model cannot find a plain sentence saying what you do, it will paraphrase your page headings, and you have no control over what it produces. Question 2, quotable answers, is also 14 points. Assistants quote complete, self contained sentences more readily than they quote fragments, so a page of marketing fragments is close to unusable as a source. Question 3, structured data, is worth 12 like most of the list, but it is the one that does not depend on wording. Organization and Product markup is how a machine confirms what you sell and who you are, independently of how your page is worded.

Questions 3 through 8 are worth 12 each, and each covers a different failure mode. Question 4 is mention volume off your own site, because a business described only by itself has a single point of failure for its own reputation. Question 5 is one claim per page, so a page can be repeated back as one idea rather than a pile of features. Question 6 is consistency of name, address, and phone, because mismatches across directories are read as a reason to be cautious about a business. Question 7 is dated, attributable content, since anything without a date and a name is hard to cite responsibly. Question 8 is your own one line pitch, which is a test of clarity that costs you nothing and is a prerequisite for the other seven answers being any good.

How this is scored, and every assumption built into it

The weights are the assumption, and they are ours, not yours. Questions 1 and 2 carry 14 points each and the other six carry 12, because we weight the two things that everything else depends on most heavily. The three options per question are worth full points, half points, and zero. Nothing in between, no partial credit for effort. Those eight weights add to exactly 100, so the total needs no adjustment and every combination shows its true sum. The bands are set at 80 and above for discoverable by AI assistants, 55 to 79 for partially readable with real gaps, 35 to 54 for mostly invisible to AI search, and under 35 for invisible. We chose those cutoffs as a readable four step ladder, not because they are derived from any measured data, and no external benchmark sits behind them. Answers are treated as accurate at face value, so if you overstate one, the score goes up for the wrong reason. The score is a snapshot of the day you take it, with no memory of your last attempt.

A worked example

Take a founder with a decent online store. They answer yes to question 1, partly to question 2, no to question 3, partly to question 4, yes to question 5, partly to question 6, no to question 7, and yes to question 8. The arithmetic is 14 plus 7 plus 0 plus 6 plus 12 plus 6 plus 0 plus 12, which is 57 out of 100. That lands in the 55 to 79 band, partially readable with real gaps. The two lowest areas are question 3 at zero of 12 and question 7 at zero of 12, so those are the two fixes the tool puts at the top. Note what that score does not say: it does not say the store will be quoted more often, appear in more AI answers, or sell more units. It says a machine currently has no structured description of the business and nothing on the site it can attribute a quote to.

The cheapest first three fixes, in order

Fix one, write the one sentence plain language description. Sit down and write a single sentence using the words your customers use, not your category's words. Put it near the top of your homepage, and use the same sentence in your meta description and your directory profiles. This takes an afternoon and nothing else, and it is the input for almost every other answer a machine gives about you.

Fix two, answer three real customer questions on a page you already have. Take the questions your support inbox and your sales DMs repeat most often, and answer each in a full sentence a stranger would understand without your context. Three questions is the whole job. This is the raw material an assistant can quote, and it usually already lives in your own message history.

Fix three, add Organization and Product structured data. One Organization block on the homepage with the name, address, and phone you settled in fix one, and Product markup on your top selling page. Keep every field identical to what the rest of your web presence already says, because inconsistent values are worse than missing ones. This is markup describing facts you have already fixed, not new copy to write.

What to do next

Work the two lowest areas your result puts at the top, in that order, then re-run this scorecard in a month. The tool costs nothing and the eight questions take two minutes, so the only thing worth deciding now is which of the two you will do this week. Questions 1 to 3 are the ones you can close yourself in an afternoon. The work that follows them, keeping descriptions, answers, markup, and directory details consistent as the business moves, is the part that quietly gets skipped once the founder is busy, which is the kind of recurring operational work we build systems for at DNA.

If you also want to know how the rest of your customer-facing surface is holding up, run the SEO Health Checker for the technical side and the Store Funnel Health Scorecard for the conversion side.