What is an AI score?
An AI score is a number that estimates how well your website performs in AI search, and every score you’ll meet belongs to one of two families that get confused constantly. Perception scores measure what AI models currently say about your brand. Readiness scores measure whether AI systems can retrieve, parse, and cite your site at all. They answer different questions, they are built from completely different measurements, and the most important thing about any AI score is not the number but whether its methodology is published, because this category has a fake-score problem.
What do the two families of AI score measure?
Perception scores work by prompting AI models about your brand and grading the answers: sentiment, share of voice, how often you’re mentioned against competitors. HubSpot’s AEO Grader is the best-known free example. These scores describe outcomes you can’t directly edit, and because model outputs are probabilistic, the same site can score differently on different days.
Readiness scores work by inspecting your actual website: can AI crawlers reach it, does the content survive being read without JavaScript, is it shaped so a model can quote it. Our AIO Score is this kind, built from 21 deterministic checks. Readiness scores describe causes you can directly fix, and the same site scores the same twice. Our free checkers roundup maps which tools give you which family.
What makes an AI score trustworthy?
One test: can you re-derive it? A trustworthy score publishes every factor and every weight, cites the evidence behind the weighting, and lets you check the math against your own page. An untrustworthy score gives you a number, a color, and a sales call.
This category is young enough that plenty of tools grade sites on factors with no measured effect, llms.txt files being the classic example, so an unverifiable score is worse than no score: it directs your effort by authority instead of evidence. The AIO Score methodology is public down to each check’s points and research precisely so it can be audited, and any score you’re paying attention to should clear that same bar.
What is a good AI score?
Against the real distribution, anything above 72 beats half the web’s biggest sites. When we scanned the web’s top 1,500 domains, the median AIO Score was 72, 45% of sites graded D or F, and only about one site in ten managed an A or better, with the full breakdown in the study.
The failures cluster in two places: blocked or firewalled AI crawlers, and content that buries its answers. Which means a good score is mostly a decision, not a talent: the checks that carry the points are all individually fixable in hours.
How do you find your AI score?
Run both families once, free, and read them together. A readiness scan tells you whether anything is stopping AI systems from seeing you, and our scan returns the full 21-check report in about fifteen seconds with a fix guide per failure. A perception check tells you what models currently say about your brand.
If perception looks bad while readiness looks good, you have a reputation and content problem. If readiness looks bad, fix that first, because no perception score can improve while the models can’t read the site it’s grading.
See where your site stands. The free scan takes about fifteen seconds and shows every fix.
Run a free AI visibility scan