What an AI visibility checker actually measures

A typical automated checker tests three groups of signals: whether named AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) can reach your pages, whether structured data such as Schema.org markup describes your business, services and location, and whether basic entity and metadata signals (title tags, meta descriptions, consistent business naming) give AI systems something concrete to read. None of these test whether an AI platform will actually mention your business; they test whether the technical groundwork for that is in place.

Checkers are not standardised the way, say, PageSpeed tests are. Each tool picks its own checklist, weights each item differently, and tests a different set of crawler user agents. One checker might flag a missing FAQPage schema as critical; another might not check FAQ schema at all. A tool that checks ten AI crawler user agents will surface blocks that a tool checking three misses entirely. Treat any single score as one data point, not a verdict, and read what was actually tested before trusting the number.

No automated checker can confirm that, and any tool suggesting otherwise is claiming more than it can measure. A checker can confirm whether your site is technically readable: crawlable, structured, described clearly. Whether an AI platform actually names your business in a generated answer depends on retrieval and ranking decisions the platform makes internally, plus what it has encountered about your business elsewhere on the web. A perfect technical score raises the odds of being considered; being included in a specific generated answer comes down to those same internal decisions.

A passing technical score means AI platforms can read your site cleanly. It does not mean they have a reason to name you over a competitor who is equally readable. When several businesses in a category are all technically sound, AI systems often lean on what they have already encountered about each one elsewhere: press coverage, review volume, mentions on other sites. That off-site layer sits outside what any on-site checker measures. QBiz Leads calls this brand distribution and citations, and it is the other half of the work a technical check cannot see.

No, and a free checker is not designed to be. A free automated check typically runs in under a minute and covers the highest-impact, most commonly broken signals: crawl access, headline schema presence, basic metadata. A full audit reviews every page individually, validates schema against the actual service list, tests multiple AI user agents against every URL pattern on the site, and produces a prioritised, page-by-page fix list. Use the free check to see whether a fuller audit is worth commissioning.

Blocked crawl access is the most damaging because it is total. If GPTBot or another AI crawler receives a 403 or a robots.txt disallow, nothing else on the page matters, since the AI system cannot read any of it. This most often comes from a security plugin's default "advanced protection" setting or a robots.txt rule written before these crawlers existed, rather than a deliberate decision. Missing schema and thin metadata are more common numerically, but a full crawl block is the single fix with the biggest before-and-after effect.

No. An llms.txt file is a plain-text summary that helps an AI system orient itself on a site quickly, and it offers convenience rather than proof of visibility. A site can publish a well-written llms.txt and still be functionally invisible if its robots.txt blocks the same crawlers the llms.txt is inviting in, or if the pages it points to are thin. llms.txt adoption remains low industry-wide (8.7% of the top 1,000 sites as of June 2026, per Rankability), so having one is a minor positive signal that most checkers do not weight heavily.

After any change to your hosting, security plugin, CDN, or a site redesign, since these are the most common sources of accidental crawl blocks. Beyond that, every three to six months is a reasonable cadence: AI platforms update how they weigh signals, and a plugin update or theme change can silently reintroduce a block that a previous check cleared.

An SEO audit tool checks ranking-relevant signals: backlinks, keyword targeting, page speed, index status, Core Web Vitals. An AI visibility checker tests a narrower, AI-specific set: whether named AI crawlers can access the site, whether structured data describes the business machine-readably, and whether content is formatted for extraction rather than for a search results snippet. A site can pass an SEO audit cleanly and fail an AI visibility check, because the two tools measure different layers of the same website.

Often, at least partially. Removing an AI crawler block from robots.txt or a security plugin's settings is usually a configuration change, not a rebuild. Adding basic Schema.org markup (Organization, LocalBusiness, Service) is achievable with a plugin or a developer's afternoon. What tends to need more sustained work is restructuring vague service pages into answer-ready content and building out FAQ sections that match real buyer questions, since that is writing and content work rather than a settings change.

A score itself is only visible to whoever runs the check on your site; it is not published anywhere public. But the underlying signals a checker reads (your schema markup, your robots.txt, your page content) are public by nature, the same way a competitor could always view your page source. A high score reflects that your site is readable, not that the readiness itself is broadcast anywhere.

You get a pass/fail read on the core signals in about thirty seconds: whether AI platforms can currently reach and read your site, which of the key AI visibility factors you are passing or failing, and a starting point for what to fix first. From there you decide whether the gaps are small enough to fix yourself or substantial enough to warrant a full audit and implementation support.