Original Research: This AI Visibility Benchmark evaluated nine leading SEO websites using the same three AI optimization tools. Every article was tested using an identical methodology, and all scores reflect the measured results at the time of testing.
AI visibility advice is everywhere. But almost nobody asks a simple question: are the websites teaching AI optimization actually following their own advice?
I decided to find out.
I picked nine well-known SEO publishers, chose one educational article from each, and ran every article through three tools from my free AI Visibility Toolkit. No manual adjustments. No editorial scoring. Just the numbers.
What I found was more interesting than I expected.
Table of Contents
Research Summary
- 9 SEO publishers tested
- 27 individual benchmark measurements
- 3 AI optimization tools
- Same methodology for every site
- Same testing period
- Original benchmark — no scores were adjusted or estimated
Benchmark at a Glance
- Highest AI Visibility score: 83 (Search Atlas)
- Highest Citability score: 91 (Neil Patel)
- Highest Schema score: 75 (Search Atlas)
- 6 of 9 tested articles scored 0 for Schema
- Average AI Visibility: 68
- Average Citability: 78
- Average Schema score: 16
That last number is the one that stayed with me. An average Schema score of 16 across nine leading SEO publishers — most of whom publish regular guidance on structured data — was not what I expected going in.
Key Findings
- Search Atlas achieved the highest overall AI Visibility score and the strongest Schema implementation.
- Neil Patel produced the most citable content in the benchmark.
- Six tested articles received a Schema score of zero, including one article specifically about schema implementation.
- High content quality alone did not guarantee high AI Visibility — several tested sites showed a significant gap between the two.
- AI Visibility and Content Citability consistently measured different things, even on the same page.
Why I Built This Benchmark
There is no shortage of AI visibility advice right now. Almost every major SEO publisher has published articles on GEO, AI citations, and AI search optimization. That advice is often good. But it is also untested in the one place that matters most — the publisher’s own site.
I wanted to see the gap between what these websites teach and what they actually implement. Not to criticize anyone. To find the patterns that any website owner could learn from. Real scores from real pages tell a clearer story than any opinion piece.
Benchmark Methodology
Why These Nine Websites
I chose nine websites that represent a broad cross-section of the SEO and digital marketing industry. All of them publish regular educational content about search optimization, AI search, or related topics. All of them have significant audiences. And, all of them have published their own guidance on AI visibility in some form.
Articles Tested
The table below shows which article was tested for each website and the primary topic it covered.
| Website | Article Tested | Topic |
|---|---|---|
| Search Atlas | How to Use AI Agents for Content Gap Analysis | AI content strategy |
| Semrush | What Is AI Sentiment Analysis? | AI search education |
| Neil Patel | How We Rebuilt AI Visibility Measurement From The Ground Up | AI visibility |
| Ahrefs | 10 ChatGPT SEO Tools | AI tools |
| Yoast | Structured Data with Schema for Search and AI | Schema and AI |
| Backlinko | How to Rank in AI Search | AI search ranking |
| Search Engine Journal | General SEO article | General SEO |
| HubSpot | Loop Marketing | Marketing education |
| Moz | Stop Measuring AI Search Like SEO | AI search strategy |
Testing Process
Every article was tested using the same three tools in the same sequence. Same day. Same methodology. One article per website. No retesting or cherry-picking of results.
The Three AI Visibility Tests
AI Visibility Checker
Measures overall AI readiness. Checks crawler access, XML sitemap, technical signals, structural elements, and metadata quality. Returns a score out of 100.
Content Citability Grader
Measures how likely a piece of content is to be cited by an AI system. Evaluates four dimensions: Evidence, Structure, Authority, and AI Readability — each worth 25 points. Inspired by publicly available GEO research factors including findings from Princeton’s study on generative engine optimization.
Schema Checker
Evaluates structured data implementation. Checks for six schema types — Article, FAQPage, Organization, WebSite, Person, and BreadcrumbList — and rates each as Complete, Basic, or Missing. Returns a score out of 100.
Overall Benchmark Results
The table below summarizes the results of all 27 benchmark measurements across nine websites and three tools.
| Website | AI Visibility | Citability | Schema |
|---|---|---|---|
| Search Atlas | 83 | 85 | 75 |
| Semrush | 78 | 77 | 35 |
| Neil Patel | 77 | 91 | 25 |
| Ahrefs | 75 | 67 | 0 |
| Yoast | 70 | 75 | 0 |
| Backlinko | 65 | 90 | 0 |
| Search Engine Journal | 65 | 59 | 0 |
| HubSpot | 61 | 83 | 10 |
| Moz | 39 | 80 | 0 |
| Average | 68 | 78 | 16 |
The results show that no publisher dominated every category. Each website demonstrated a different combination of strengths and weaknesses, making the benchmark more nuanced than a simple ranking.
AI Visibility Hall of Fame
🏆 Best Overall AI Visibility — Search Atlas (83) Highest AI Visibility score in the benchmark. Strong across every measured dimension.
🏆 Best Content Citability — Neil Patel (91) Highest citation potential of any tested article. Exceptional educational structure and evidence quality.
🏆 Best Schema Implementation — Search Atlas (75) Only website to score above 35 on schema. Semrush came second at 35. Six of nine tested articles scored zero.
🏆 Most Balanced Performer — Search Atlas Only website to rank near the top in all three categories simultaneously. No obvious weakness relative to the other tested sites.
🏆 Biggest Surprise — Yoast The tested article was specifically about structured data and schema for AI search. The Schema Checker did not detect the measured schema types on that page. That gap between topic and implementation was the single most unexpected result in the benchmark.
🏆 Most Interesting Outlier — Ahrefs A Schema score of zero combined with an AI Visibility score of 75. The clearest data point suggesting that AI Visibility does not depend entirely on structured data.
🏆 Citation Champion — Neil Patel Highest Citability score at 91. Backlinko was close behind at 90. Both produced content that scored exceptionally well on evidence, educational structure, and authority signals.
Overall AI Visibility Scores
AI Visibility Checker results — 9 leading SEO websites tested with identical methodology
Content Citability Scores
Content Citability Grader results — educational quality varied more than expected
Individual Website Analysis
Search Atlas
AI Visibility: 83 | Citability: 85 | Schema: 75
Search Atlas was the only website to rank near the top across all three measurements. Its AI Visibility score led the benchmark. Its Citability score was second only to Neil Patel. Its Schema score was the highest by a significant margin — 75 compared to Semrush’s 35 in second place.
The tested article combined technical signals with high-quality educational content. The schema implementation was the most complete of any site in the benchmark.
Research conclusion: Search Atlas demonstrated the most balanced AI optimization profile in this benchmark.
Semrush
AI Visibility: 78 | Citability: 77 | Schema: 35
Semrush scored second in AI Visibility and produced balanced scores across all three tools. Its Schema score of 35 was the second highest — well ahead of most competitors, though significantly behind Search Atlas. Content Citability was solid but not among the leaders.
Not outstanding in any single category, but consistent across all three. That consistency matters.
Research conclusion: Semrush showed one of the most technically balanced implementations in the benchmark.
Neil Patel
AI Visibility: 77 | Citability: 91 | Schema: 25
Neil Patel produced the highest Content Citability score in the benchmark at 91. The tested article — about rebuilding AI visibility measurement — had exceptional educational structure, strong evidence signals, and clear first-hand authority. AI Visibility was strong at 77. Schema implementation was moderate at 25.
The gap between Citability and Schema was notable. This was content that AI systems would find highly citable on its merits, even without strong structured data support.
Research conclusion: Neil Patel produced the benchmark’s most citation-friendly content.
Ahrefs
AI Visibility: 75 | Citability: 67 | Schema: 0
Ahrefs scored 75 on AI Visibility despite a Schema score of zero — the clearest example in the benchmark of strong AI Visibility without detectable structured data. Citability at 67 was lower than expected given the brand’s authority and content volume.
That combination — strong technical AI signals, no detected schema, moderate citability — suggests AI Visibility draws on a wider range of signals than schema alone.
Research conclusion: AI Visibility does not appear to depend entirely on structured data, based on this tested article.
Yoast
AI Visibility: 70 | Citability: 75 | Schema: 0
Yoast was the biggest surprise in the benchmark. The tested article was specifically about structured data and schema for search and AI. The Schema Checker did not detect the measured schema types on that page.
AI Visibility was reasonable at 70. Citability was solid at 75. But the absence of detectable schema on an article explicitly about schema implementation was the most striking single finding in the benchmark.
Research conclusion: Technical expertise in a topic did not translate into schema implementation on the tested page.
Backlinko
AI Visibility: 65 | Citability: 90 | Schema: 0
Backlinko produced the second-highest Citability score at 90 — just one point behind Neil Patel. The tested article on ranking in AI search had exceptional educational quality and very strong citation potential. AI Visibility at 65 was mid-range. The Schema Checker detected no schema on the tested page.
The content quality was outstanding. According to the benchmark, the tested page showed fewer technical optimization signals than several competing articles.
Research conclusion: Excellent educational content can achieve very high Citability scores without detectable structured data.
Search Engine Journal
AI Visibility: 65 | Citability: 59 | Schema: 0
Search Engine Journal scored mid-range on AI Visibility and had the lowest Content Citability of any tested publisher at 59. The Schema Checker detected no schema on the tested page. The article appeared to rely more on brand authority than educational optimization signals.
Brand authority and publishing volume did not appear to translate into stronger AI citation potential on the tested page.
Research conclusion: Strong brand authority does not necessarily correspond to stronger AI citation potential based on this benchmark.
HubSpot
AI Visibility: 61 | Citability: 83 | Schema: 10
HubSpot had the largest gap between Citability and AI Visibility in the benchmark. Content quality was very high at 83. Technical AI Visibility was lower at 61. Schema implementation was minimal at 10.
The educational quality of the tested article was well above the benchmark average. According to the benchmark, the tested page showed fewer technical optimization signals than several competing articles.
Research conclusion: Content quality substantially exceeded technical optimization signals on the tested article.
Moz
AI Visibility: 39 | Citability: 80 | Schema: 0
Moz had the lowest AI Visibility score in the benchmark at 39 — and the largest absolute gap between AI Visibility and Citability of any tested site. The tested article scored 80 on Citability, placing it above the benchmark midpoint. The Schema Checker detected no schema on the tested page.
The article itself had strong educational content and clear authority signals. According to the benchmark, the tested page showed fewer technical optimization signals than several competing articles — a gap that the content quality alone could not close.
Research conclusion: Publishing excellent educational content alone did not guarantee strong AI Visibility on the tested page.
Patterns I Discovered
These are the findings I found most useful — not conclusions about individual companies, but patterns that showed up consistently across the nine tested pages.
The table below shows the key observations at a glance.
| Pattern | Observation |
|---|---|
| Highest AI Visibility | Search Atlas (83) |
| Highest Citability | Neil Patel (91) |
| Highest Schema | Search Atlas (75) |
| Lowest AI Visibility | Moz (39) |
| Average Schema score | 16 / 100 |
| Pages with zero schema | 6 of 9 |
Content quality and AI Visibility measured different things
The correlation between Citability and AI Visibility was weaker than I expected. Neil Patel scored 91 on Citability and 77 on AI Visibility. Backlinko scored 90 on Citability and 65 on AI Visibility. Moz scored 80 on Citability and 39 on AI Visibility.
High educational quality did not automatically produce high AI Visibility. The gap was large enough on several tested pages to matter. These are genuinely different signals.
Schema was not the strongest predictor of AI Visibility
Ahrefs scored 75 on AI Visibility with a Schema score of zero. Backlinko scored 65 with zero schema. Yoast scored 70 with zero schema. Structured data appeared to contribute to AI Visibility but was not a requirement for it. Other signals — technical accessibility, content structure, metadata quality — appeared to carry weight even in the absence of JSON-LD schema.
The average Schema score across all nine tested articles was 16. That is a low number for a group of publishers that collectively produce some of the most-read AI optimization content on the web.
Schema Implementation Scores
Schema Checker results — 6 of 9 tested articles received a score of zero
The highest-performing websites combined multiple strengths
Search Atlas was the clearest example. It did not have the highest Citability score — Neil Patel did. It did not have a perfect Schema score. What it had was consistent strength across all three measurements. That balance appeared more valuable than excelling in one area while lagging in others.
Large brands did not automatically score highest
Several publishers with significant domain authority and large content libraries scored lower than expected in one or more categories. Brand authority and publishing frequency did not predict AI optimization scores on the tested articles.
Every publisher had a different optimization profile
No two websites showed the same pattern of strengths and weaknesses. Some prioritized content quality. Some had stronger technical signals. And some had neither in a strong combination. There is no single dominant approach to AI optimization among the tested publishers.
What Website Owners Can Learn
The lessons from this benchmark are consistent regardless of the size of the website you run.
Getting the fundamentals right produced the strongest results. AI crawler access, complete schema, clear heading structure, specific evidence in content, and genuine author signals were the factors that separated the stronger performers from the weaker ones. None of those are expensive or complex to implement.
Do not chase a single metric. The websites with the highest overall profiles were not perfect on any single tool. They were consistently good across all three. A Citability score of 90 with an AI Visibility score of 39 represents a real gap that content quality alone cannot close.
Schema matters more than most publishers in this benchmark were treating it. Six of nine tested pages returned a Schema score of zero. The average across all nine was 16. That gap between advice and implementation was one of the clearest findings in the data.
Content quality is a genuine signal. Neil Patel and Backlinko scored 91 and 90 on Citability respectively. That reflects real educational structure, real evidence, and real authority signals. Those scores are achievable for any publisher willing to invest in original, specific, well-structured content.
Limitations
This benchmark has real limitations worth stating clearly.
Only one article per website was tested. These scores reflect the tested pages — not the overall optimization level of each publisher’s entire site.
The benchmark reflects scores at a single point in time. Websites update content and technical implementation regularly. Scores will change.
The toolkit measures practical AI visibility signals. It does not predict future citation rates, which depend on specific prompts, model versions, and factors outside the scope of any current tool.
These results should be read as a comparative snapshot of nine tested articles on a given date — not as definitive rankings of the companies behind them.
Nena’s Quick Verdict
The biggest surprise was not which website scored highest.
It was how different the optimization strategies were across nine publishers who all write about the same topic. Some excelled through exceptional educational content. Some had stronger technical signals. Very few combined both at a high level.
Search Atlas demonstrated the most balanced profile in this benchmark. Neil Patel and Backlinko produced the most citation-friendly content. Moz showed the largest gap between content quality and technical AI readiness. And the website whose tested article was specifically about schema returned a Schema score of zero.
The average Schema score across nine leading SEO publishers was 16. That single number tells you more about the current state of AI optimization than most articles I have read on the topic.
For website owners, the message is clear. AI visibility is not one thing you optimize. It is a combination of technical signals, content quality, and structured data working together. The strongest tested pages were not perfect on any single measure. They were consistently good across all of them.
That is harder to achieve than chasing one metric. It is also more durable.
Frequently Asked Questions
AI visibility is the degree to which an AI system can discover, understand, and choose to reference a website when generating a response. It depends on crawler access, schema markup, content structure, and metadata signals like title tags and sitemaps.
The AI Visibility Checker evaluates ten signals across website health and page health and returns a score out of 100. It checks crawler access, sitemap, schema, title tags, meta descriptions, headings, and content depth.
A content citability score measures how likely a piece of content is to be cited by an AI system. The Content Citability Grader evaluates evidence, structure, authority, and AI readability — each worth 25 points for a total score out of 100.
Schema tells AI systems what a page is, who wrote it, and what it covers. Complete schema removes ambiguity and makes content easier to classify and cite accurately.
Based on this benchmark, yes. Ahrefs scored 75 on AI Visibility with a Schema score of zero. Schema helps but other signals — crawler access, content structure, metadata — carry meaningful weight in its absence.
GEO is the practice of optimizing content to appear in AI-generated responses from ChatGPT, Claude, Perplexity, and Google AI Overviews. It draws on research including findings from Princeton’s study on AI citation factors.
The strongest signals are: specific statistics, clear heading structure, named authorship, first-person experience language, explicit definitions, FAQ sections, and short paragraphs. Pages that combine several of these signals consistently score higher.
