AI Visibility Score Explained: What the Numbers Actually Mean

The AI Visibility Checker gives you three scores and a grade, and nine out of ten people look at the top-right number and stop. I get it — the big grade is satisfying. But it's also the least useful thing on the screen. The value lives in the two component scores and the engine-by-engine detail, because those point to a specific problem instead of just a number.

Let me decode each piece, and which one you should actually act on.

The website score

This is the foundation. It measures how ready the site itself is to be understood and trusted. Think indexability, structured data, content depth, security, trust signals, readability, mobile experience. It answers the question: "is my website the kind of thing an AI would want to cite?"

A high website score means the groundwork is good. But it's a measure of readiness, not outcome. You can be fully ready and still invisible, which is why the website score on its own can be misleadingly reassuring.

The website score evaluates several concrete signals:

• Indexability — Can search engines and AI crawlers access and read your pages? Blocked resources, noindex tags, and crawl errors all lower this score.
• Structured data — Do you have the schema markup that helps engines understand your entity, services, and content? Missing Organization, LocalBusiness, or FAQPage schema creates blind spots.
• Content depth — Are your pages substantial enough to extract meaningful facts? Thin pages with generic advice score poorly because there's nothing specific for an AI to cite.
• Trust signals — Do you have reviews, case studies, author information, and external validation? Trust signals tell engines your content is worth recommending.
• Technical health — HTTPS, mobile responsiveness, page speed, and core web vitals all factor into whether engines consider your site reliable.

If your website score is low, the full AI visibility audit walks through fixing each of these areas.

The AI score

This is also the number our ChatGPT recommendation tracking plots for you week over week, so you watch it move instead of wondering.

This is the outcome. It measures whether AI engines actually surface and cite you — not whether you deserve to be, but whether they do. When it's low, the engines aren't choosing you as their source, picking a competitor instead, or not mentioning you at all.

This is the number that matters most for visibility, and it's the one you can't see anywhere else. Google Search Console shows you clicks and impressions from traditional results. It won't tell you that ChatGPT recommended a competitor in its answer.

The AI score is calculated by running real buyer-intent prompts across multiple AI engines and measuring:

• Citation rate — How often do AI engines cite your website as a source when answering relevant questions?
• Recommendation rate — How often do they actively recommend your business by name?
• Accuracy — When they do mention you, is the information correct and current?
• Consistency — Do the engines agree about who you are, or are they working from conflicting sources?

A low AI score with a high website score is the most common scenario we see. It means you've built a technically sound site, but AI engines aren't connecting the dots between your content and buyer-intent questions. The fix isn't more technical work — it's improving your entity signals, content extractability, and cross-web consistency.

The gap between them is the insight

Here's the move most people miss: don't read the two scores separately. Read the gap.

• High website, high AI — You're in good shape. Maintain it and keep your info consistent. Focus on monitoring and incremental improvements.
• High website, low AI — This is the classic case. Your site is technically healthy but AI isn't citing you, which usually points to an attribution problem: inconsistent business details, thin content, or missing entity clarity. This is very fixable, and the fail checks show you where.
• Low website, low AI — Start with the technical and content foundation first. Fix the fails, then worry about getting cited. Trying to earn AI citations with a broken site is like advertising a store with no sign.
• Low website, high AI — Rare, but it happens when third-party mentions and directory listings carry your visibility while your own site lags. Fix the technical gaps before the external signals dry up.

Understanding this gap is the difference between random fixes and a focused strategy. Our guide on why your AI visibility score might be low digs deeper into each scenario.

What the grade and checks add

The overall grade is a useful summary for that weekly report you're showing someone else. But for your own decisions, work from the checks. Each one is a pass, warn or fail on a concrete signal. The fails are your to-do list, in priority order. Fix them, rerun, and watch them flip.

The checks break down into categories:

• Technical checks — Indexability, structured data, HTTPS, mobile experience. These are binary: either the signal is present or it isn't.
• Content checks — Depth, extractability, originality, freshness. These are gradients: your content can be good but not great, or great but not optimized for extraction.
• Trust checks — Reviews, citations, external mentions, author credibility. These build over time and require ongoing effort.
• Consistency checks — Business information alignment across your site and third-party sources. One inconsistency can undermine otherwise strong signals.

Don't try to fix everything simultaneously. Focus on the fails, because they represent the biggest gaps between your current state and what AI engines need to trust you.

Reading the engine-by-engine view

One score hides a lot. Your brand might be loved by ChatGPT and invisible to Perplexity. So look at each engine separately.

When an engine matches you on your name, it mentioned your business but might be drawing on a directory listing or a review site. When it matches you on your domain, it cited an actual page from your site — the stronger signal, because it means the AI trusted your own content.

Consistency matters here too. If three engines describe you one way and a fourth has a stale, different version, that inconsistency is a problem in itself. AI engines blend sources, and contradictory facts make you harder to trust across all of them.

Different engines have different source preferences, which explains why your performance varies:

• ChatGPT tends to favor authoritative, well-structured content with clear entity signals and strong domain authority.
• Perplexity weights recent content heavily and pulls from a broader web index, making freshness and third-party citations important.
• Gemini leverages Google's own index and entity graph, so traditional SEO signals like Google Business Profile optimization carry extra weight.
• Claude focuses on content quality and depth, often favoring comprehensive, well-argued pages over thin, keyword-stuffed ones.

Understanding these patterns helps you prioritize. If you're invisible on Perplexity, build more third-party citations. If Gemini doesn't surface you, improve your Google Business Profile and traditional SEO signals.

Our guide on how to rank in ChatGPT search goes deeper into engine-specific optimization strategies.

What a high website score with a low AI score actually means

This is the scenario we see most often, and it's worth unpacking in detail because it's the one where businesses waste the most time. The website score says 75. The AI score says 30. The grade lands in the middle and tells you nothing useful.

What's happening: the site itself is fine. It loads quickly, has structured data, has decent content. But AI engines aren't picking it up because the signals that connect your site to your business identity are weak or scattered.

The most common culprits:

1. Business information is inconsistent across the web. Your GBP says one thing, your website says another, and a directory listing has a third version. AI engines can't confidently identify you as a single entity. Our AI trust signals guide explains why consistency is the foundational requirement.

2. Content is well-written but not extractable. The pages read well for humans but bury their key facts. An AI engine looking for "what services does this business offer" has to parse three paragraphs to find the answer — a competitor who lists services clearly wins the citation.

3. No third-party corroboration. Your site makes claims that no independent source backs up. AI engines want to verify before they recommend, and a site with zero external mentions is unverifiable.

4. Category mismatch. Your Google Business Profile says one category, your website content implies another, and your schema declares a third. This fragmentation makes it harder for engines to match you to the right queries.

The fix isn't to overhaul the website — it's to fix the attribution and consistency layer. Make the business identity unmistakable across every signal. The website score stays the same; the AI score climbs.

How to act on the scores

Here's the prioritised action sequence:

1. If your website score is low: Fix the technical foundation first. Add missing schema, fix crawl errors, improve page speed, ensure mobile responsiveness. Our structured data for AI visibility guide covers the schema implementation.

2. If the gap is large (high website, low AI): Focus on entity consistency, content extractability, and third-party mentions. This is where the biggest ROI lives.

3. If your AI score varies wildly by engine: Build a balanced presence. Improve the weakest engine's signals by targeting the specific source preferences that engine relies on.

4. After any fix, re-run in 30 days. Track the trend. A score that moves from 30 to 45 after fixing consistency tells you the strategy is working. A score that doesn't move tells you to look at the next check.

The short version

Stop reading the grade and start reading the story. Website score tells you if your site is ready. AI score tells you if AI is actually choosing you. The gap tells you what to fix. Engine by engine tells you where you're strong and where you're ignored.

The most actionable approach: run the check, note the gap between website and AI scores, fix the failed checks, then re-run in 30 days. Track the trend, not the snapshot. AI visibility improves with consistent, focused effort — not with one-time fixes.

For a deeper dive, see our guide on what AI visibility actually means.

For a deeper dive, see our guide on the AI visibility checker guide.

Want a hand reading yours? book a free website teardown at Launch at Dawn and we'll decode your report with you and turn it into a fix list.