AI SEO Analyzers: What Their Scores Mean and Which Ones to Trust

99Visibility Research · Updated October 21, 2026 · 8 min read

Short answer

An AI SEO analyzer crawls a page and scores how ready it looks for AI search. Many of those scores lean on factors primary sources reject, such as schema for citations, content chunking and an "AI-friendly" writing style. The checks worth trusting report access facts: whether each AI crawler may fetch the page, whether a firewall blocks it, whether key content exists before JavaScript runs, and whether snippet controls hide it.

Key takeaways

An AI SEO analyzer score blends mechanical facts about your site with opinions about what AI engines prefer, so read the line items, not the total.
Schema for citations, chunking, AI-friendly style, llms.txt, E-E-A-T scores and domain authority have no credible evidence as levers for AI answers.
The checks that hold up are access checks: per-bot robots.txt, firewall and CDN reachability, raw HTML versus rendered content, and snippet controls.
No page analyzer can see the queries engines actually run or where your page sits in a model's context, so no analyzer can promise citations.
Fix blocks first, then state your key facts plainly in fetchable HTML, then measure each AI platform separately.

What an AI SEO analyzer actually scores

Most AI SEO analyzers fetch one URL, run a list of checks, and roll the results into a score out of 100. The checks usually fall into five groups:

Access: robots.txt rules per AI crawler, HTTP status codes, firewall or CDN responses.
Rendering: whether the main content is in the raw HTML or only appears after JavaScript runs.
Markup: schema types, FAQ markup, llms.txt presence, meta tags.
Structure and style: heading counts, paragraph length, "chunkability", reading level, question-style headings.
Authority: an E-E-A-T rating, author signals, a domain authority number.

The first two groups describe mechanical facts about your site. You can verify them, and they have a documented consequence. The last three mostly describe opinions about what AI engines prefer.

A single total blends both kinds. So a page can score 85 while OAI-SearchBot, the crawler behind ChatGPT search, gets a 403 from your firewall. The number looks healthy and the page is invisible to that engine. Treat the total as a summary and the line items as the real report. If you want to see how a single page fares before buying anything, a free AI SEO checker covers the basics in minutes.

How AI engines use a page, and what an analyzer can see

An AI engine uses your page in four stages. Separating them shows what any analyzer can and cannot measure.

Parametric: what the model already knows from training, frozen at a knowledge cutoff. Nothing you change this quarter reaches it.
Access: whether a crawler may fetch and store the page at all. You and your CDN control this.
Retrieval: whether the page enters the model's context for a given query. The engine controls this.
Generation: which retrieved pages the model uses and cites. The model controls this.

A page analyzer can only test stage 2 with confidence. It can request your page, read your robots.txt and compare raw HTML with rendered HTML.

Stages 3 and 4 are mostly hidden. OpenAI says ChatGPT search rewrites a prompt into one or more targeted queries, and Google describes a query fan-out across subtopics. Nobody outside the engine sees those queries.

The research matches this. A July 2026 critical survey of 45 GEO studies (arXiv:2607.14035) found that already-retrieved content can causally alter its citation or use, but no reviewed technique shows a stable, longitudinal, cross-platform causal effect on discoverability. Nearly every published GEO effect was measured on pages already inside the model's context. An analyzer that scores your "citation readiness" is scoring a stage it cannot observe.

Which analyzer scores rest on rejected factors

Several common line items rest on claims that primary sources contradict. Use this table to read an analyzer report.

Analyzer check	What the evidence says
Schema "for AI citations"	Google says structured data isn't required for generative AI search. Ahrefs tracked 1,885 pages adding schema against matched controls and found AI Overview citations fell 4.6%.
FAQ schema	Google deprecated FAQ rich results in May 2026.
Chunking or passage score	Google's AI optimization guide rejects chunking as a requirement and says there is no ideal page length.
"AI-friendly" style, long-tail coverage	Google says you don't need to write in a specific way for generative AI search or capture every keyword variation.
llms.txt present	Google says Search ignores it. No other engine documents reading it.
E-E-A-T score	Google says E-E-A-T isn't a specific ranking factor. There is no score to measure.
Domain authority	A third-party vendor metric. No engine computes it (see what domain authority is and isn't).
Freshness score	Dates are self-declared, and no study shows engines reward them.

Don't overcorrect. Google says existing markup does no harm, so ripping out schema is not supported either. The honest line is narrower: schema is not a lever for AI answers. Ahrefs' treated pages were already heavily cited, so whether schema helps a new page enter an answer is untested. Untested is not a reason to pay for it or rank it highly in a score.

The checks worth trusting

These checks report facts with a consequence documented by the engine operators. An analyzer that leads with them is telling you something useful.

Per-bot robots.txt permission

robots.txt rules apply bot by bot. OpenAI's crawler page says each setting is independent: GPTBot governs training, OAI-SearchBot governs ChatGPT search. Perplexity uses PerplexityBot, Anthropic uses Claude-SearchBot, and Meta uses Meta-WebIndexer for Meta AI search. A good analyzer lists each bot and the consequence of blocking it.

Network reachability

A firewall or CDN can block a crawler that robots.txt allows. OpenAI's help centre says eligibility needs both: allow OAI-SearchBot, and confirm your host or CDN allows OpenAI's published searchbot IP addresses. The only real test is to request the page as each crawler and record the status code.

Rendering

Vercel found that ChatGPT and Claude crawlers fetch JavaScript files (11.50% and 23.84% of requests) but don't execute them. Content that appears only after JavaScript runs may be invisible to those crawlers. Googlebot does render. The check is a diff: raw HTML against the rendered page.

Snippet and index controls

Google and Bing document noindex, nosnippet, max-snippet and data-nosnippet as controls that keep content out of AI answers. These often get inherited from a template by accident. A good analyzer flags every instance.

What is in your control and what is not

In your control:

Which crawlers your robots.txt allows, per bot.
Whether your firewall, CDN and bot rules let those crawlers through.
Whether your prices, product names and key claims sit in the HTML before JavaScript runs.
Whether snippet directives hide content you want shown.
Whether your own pages state the same fact the same way. A specific, fetchable sentence (a price, a number, a one-line definition) gives an engine something it can quote.

Not in your control:

The queries the engine actually runs.
Where your page lands in the model's context.
Whether a given answer used live retrieval at all.
The model's training cutoff.

This is why page scores and brand scores differ. An analyzer scores one page. A visibility score samples AI answers about your brand, and it averages answers given with and without retrieval. Read how an AI visibility score is calculated before you report either number upward. When your tools disagree, the cause is usually different prompts, engines and retrieval states, which why AI visibility tools disagree breaks down.

A practical order of work

Work from blocks to facts. Each step removes a reason an engine cannot use your page.

Read your robots.txt per bot. Confirm you treat training and search bots on purpose. A common setup blocks training while allowing search:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

Test the firewall. Request key pages as each AI crawler and record the status codes. The free AI Crawler Checker on 99visibility.com shows which crawlers your firewall turns away. If you use Cloudflare, review your AI crawler settings: from 2026-09-15 it blocks Training and Agent crawlers by default on ad-displaying pages for new domains and free-tier customers.
Diff raw and rendered HTML. List every price, plan name and claim that appears only after JavaScript runs. Move it into server-rendered HTML.
Search for suppression. Find noindex, nosnippet, max-snippet:0 and data-nosnippet, and confirm each one is intentional.
Check index status engine by engine. Google Search Console and Bing Webmaster Tools report separately. One healthy index tells you nothing about the other.
Fix contradictions. Make your pricing, about page and footer state each fact once, the same way.
Measure each platform separately. A per-platform view shows which engine names you and which does not.

Common mistakes when reading analyzer reports

Treating the total as a forecast. A high page score does not predict citations. No analyzer sees retrieval or generation.
Allowing GPTBot to get into ChatGPT search. GPTBot governs training only. Search uses OAI-SearchBot.
Blanket-blocking every AI bot. A single rule meant to stop training can remove you from ChatGPT, Perplexity and Claude search at once.
Trusting robots.txt alone. A 403 from your firewall overrides an allow rule, and nothing in robots.txt shows it.
Rewriting pages to chase a style score. C-SEO Bench (NeurIPS 2025) found most conversational SEO methods largely ineffective and often harmful to a document's ranking.
Adding quotes and statistics for show. Appended quotation blocks and injected statistics are tactics the evidence flags as risk, not wins.
Using hidden text or instructions aimed at AI crawlers. Injected content and cloaking are illegitimate. Research shows some search agents can be manipulated, which is a reason to watch third-party pages about your brand, never a tactic to use.

If you're considering an automated tool that also edits pages, read the risks of AI SEO agents first.

Frequently asked questions

Does a high AI SEO analyzer score mean AI will cite my page?

No. An analyzer can confirm your page is reachable, rendered and not suppressed. Those are conditions for being used, not guarantees. Retrieval and citation happen inside the engine, on queries you cannot see. A high score from markup or style checks says even less, because those factors lack evidence as levers. Treat a clean access report as removing blocks, and measure actual AI answers separately.

Should I remove schema because analyzers overrate it?

No. Google says existing structured data does no harm, and schema still serves its documented purposes in classic search features. The evidence only says schema is not a lever for AI citations: Ahrefs' matched test found AI Overview citations fell 4.6% after pages added it. Keep what you have, stop paying for schema work sold as an AI fix, and don't let it inflate a readiness score.

Can an analyzer tell if my firewall blocks AI crawlers?

Only if it actually requests your page as each crawler and records the response. Many analyzers just read robots.txt, which cannot show a firewall or CDN block. Look for a tool that sends requests with each crawler's user agent and reports status codes such as 200, 403 or a challenge page. Your own server or CDN logs, filtered to verified bots, are the ground truth.

Why do two AI SEO analyzers give my page different scores?

Each tool picks its own checks and weights, and many include factors with no evidence behind them. One may weight schema heavily while another weights rendering. The totals are each vendor's model, not a measurement of how engines treat you. Compare the individual line items instead. Access and rendering findings should match across tools, because they describe the same mechanical facts.

Is an llms.txt check worth anything in an analyzer report?

It tells you whether the file exists, nothing more. Google says Search ignores llms.txt and that creating one neither helps nor harms visibility in Google Search. No other major engine documents reading it. Keeping one for other services is fine, but an analyzer that adds points for it is rewarding a file with no documented effect on AI answers.

Sources

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.