AI Search Readiness | Scoped SEO Docs

What Scoped SEO checks for AI search, how the score is calculated, and what it deliberately does not claim.

This is the tab labelled AI Visibility in the sidebar. The score it produces is called AI Search Readiness.

AI Visibility tab showing a 100 out of 100 AI Search Readiness score, a Strong readiness signals badge and the AI crawler access table — The AI Visibility Overview sub-tab: the score, a tier badge, check counts and the start of the crawler table.

The tab has four sub-tabs: Overview (the score, a tier badge, how many checks passed or need review, and the crawler table), Crawler Access, Google AI and Category Checks.

What the score means. AI Search Readiness evaluates observable technical and content signals that may support search discovery and understanding. It does not measure actual AI citations or guarantee visibility.

What is checked

Crawler access from robots.txt

Each named crawler is evaluated against the page's own robots.txt, for this page's path, with the same RFC 9309 matching rules used on the Indexability tab.

Group	Crawlers	In the score?
Search crawlers	OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot	Yes, 4 points each
Training crawlers	GPTBot, ClaudeBot	Never. Shown separately, and blocking them never changes the score
AI-use control	Google-Extended	Never. Shown as an AI-use control that does not remove a page from Google Search
User-initiated fetchers	ChatGPT-User, Claude-User, Perplexity-User	Never. These fetch on a person’s request and operators say they may not follow robots.txt

Each crawler row shows its purpose and a status, and expands to the deciding group and rule, a confidence level and the operator's own caveat. Statuses are Allowed by evaluated robots.txt rules, Disallowed by evaluated robots.txt rules, Unable to verify and Not checked. When a * group is what blocks a crawler, a button copies an example rule. A collapsed list covers other AI-related tokens.

Heads up. Statuses describe what robots.txt asks a crawler to do. A CDN, firewall or login can still block a crawler, and Scoped SEO never sends a request using a crawler’s user agent, so it can’t prove a real crawler can or can’t reach the page.

Technical accessibility, content clarity and entity information

Main text in the initial HTML. The visible main text is compared with the HTML the browser received, using a repeat request from your browser. It compares word counts, not meaning. A bot-protected site may answer the repeat differently, in which case the check reads Not evaluated.
Indexing directives: response status, noindex, snippet limits and canonical.
Readable text, title, headings in longer text and page language.
Entity markup. Only judged when you declare Organization or Person markup. Otherwise it reads Not applicable.
Informational only, no weight: entity evidence, attribution on article-type pages, links, llms.txt availability, access heuristics and noai tags.

How the score is calculated

The model is called ai-score/2 and is documented in the tab under "How this score is calculated".

Category	Points	Checks
Technical accessibility	50	Response status 6, noindex 10, snippet limits 4, four search crawlers 4 each, canonical 4, main text in initial HTML 10
Content clarity	30	Readable text 12, title 8, headings in longer text 7, page language 3
Entity information	20	Entity name 10, entity URL 10 (only when markup is declared)

A pass earns full weight, a limited result earns half, a fail earns none.
Not applicable and Not evaluated checks are left out. Not evaluated also lowers coverage.
With under 50% coverage, fewer than three evaluated checks, or no technical check evaluated, the score reads Not evaluated with no number.

Tiers

A tier badge sits beside the ring. It is a label, not a prediction: Restrictions found (any verified restriction), Needs attention (under 65), Partly evaluated, Some gaps found, Strong readiness signals and Not evaluated.

What is never penalised

Missing FAQ schema, no llms.txt, non-question headings, no answer-first opening, keyword density, templates, AI-specific markup, blocking training crawlers and non-English pages. llms.txt is informational only, and the tab notes that Google does not use it for Search.

Google AI features section

A separate section reuses the Indexability results for Google's AI features and links to Google's guidance. Search Console controls for generative AI can't be read from a page, so they are listed as Not checked.

Limits to keep in mind

No live crawler testing. Nothing proves a real crawler can or can't reach a page.
A partly evaluated scan can show a high number over fewer checks. The badge, ring colour and crawler rows say when it is partial.
Question-heading detection is English only and says so.
Crawler names and purposes follow operators' public documentation at the time of the release.
Scores from 2.14 on are not comparable with the earlier "AI Visibility" pass rate. A page can move either way with no change to the site. Saved audits are never rescored.