AI Search Readiness Report
Fetch one public page for three disclosed heuristic reviews: AI search structure, GEO signals and visible E-E-A-T evidence.
One fetch. Three transparent perspectives.
PUBLIC PAGE / BOUNDED CHECKGood input. Useful output.
Fetch one public page for three disclosed heuristic reviews: AI search structure, GEO signals and visible E-E-A-T evidence.
Try the example, or make it your own.How to use the
AI Search Readiness Report
- 01
Enter the final public page URL
Use a public HTTP or HTTPS page without credentials. Redirect hops and each destination’s robots policy are validated.
- 02
Review all three heuristic sections
One returned HTML document feeds AI search structure, GEO and visible E-E-A-T checklists. Scores disclose their checks and weights.
- 03
Fix the underlying evidence
Improve missing metadata, headings, sources and identity information, then rerun. No section measures rankings, citations or real expertise.
Eight checks, with 14 possible weight units
The percentage is passed weight divided by observed weight, rounded to a whole number. The 150-word condition is a tool heuristic, not an AI-provider requirement or a minimum amount of content that every page should contain. Do not pad a useful short answer to satisfy it.
| Check | Passing condition | Weight |
|---|---|---|
| Extractable page text | At least 150 words after removing navigation and other non-content elements. | 2 |
| Clear heading structure | Exactly one H1 and at least one H2. | 2 |
| Page context | Both title and description are present. | 1 |
| Parseable structured data | At least one JSON-LD block parses and none fail parsing. | 2 |
| Named author | Recognized author metadata or a JSON-LD author is found. | 2 |
| Publication context | Recognized publication-date metadata is found. | 1 |
| External references | At least one external HTTP(S) link exists. | 2 |
| About / contact references | A recognized about or contact path is linked. | 2 |
A higher score should reflect a better explanation
Consider a guide that passes the content, headings, page-context and structured-data checks but has none of the four identity/reference signals. It earns seven of fourteen weight units. Adding a genuine author and relevant source links would add four units, if the parser can detect them.
Passing weights: 2 + 2 + 1 + 2 = 7
Observed weights: 14
Checklist score: 50/100After real author and reference improvements:
Passing weights: 11 of 14
Checklist score: 79/100Presence is not quality, and HTML is not the whole experience
The structured-data check parses JSON syntax. It does not validate Schema.org vocabulary, rich-result eligibility or whether the claims in the markup are true. The external-reference check counts links without judging their authority, relevance or support for the surrounding statement.
The fetcher reads server HTML rather than executing a browser. A visible byline rendered only in JavaScript may be missed. Navigation, header, footer, scripts and similar elements are excluded from the main-text word estimate. Compare the actual returned source when a finding differs from your visual impression.
Use the audit as an editorial checklist
Answer a concrete question early
A clear definition, procedure or conclusion helps a reader understand the page's purpose. This tool does not assess answer correctness, so editorial review still owns that work.
Make authorship meaningful
A name is only a signal. Connect it to relevant experience, a useful biography and a clear publisher identity when the content warrants it.
Cite evidence where a claim needs it
A random outbound link can satisfy the mechanical check without helping anyone. Choose a source that supports the particular statement and explain the connection.
Keep structured facts aligned
Dates, author details and page entities should agree with the visible content. Do not change dates merely to create an appearance of freshness.
The GEO scoring method: nine checks, 15 weight units
These weights belong to Outgle's checklist. They are not published ranking weights from Google or any AI provider. The presence tests are deliberately inspectable, but they cannot measure the quality of an author, the credibility of a source or the truth of a statement.
| Signal | What earns the points | Weight |
|---|---|---|
| Extractable text | At least 150 words in the filtered server HTML. | 2 |
| Heading structure | One H1 and at least one H2. | 2 |
| Page context | A title and a meta description are present. | 1 |
| JSON-LD parsing | One or more JSON-LD blocks parse, with no invalid blocks. | 2 |
| Authorship | Recognized author metadata or JSON-LD author value. | 2 |
| Publication context | Recognized publication-date metadata. | 1 |
| External references | One or more external links. | 2 |
| Identity links | A linked about or contact path matching the implemented pattern. | 2 |
| Crawler access | OAI-SearchBot is allowed for the exact path in robots.txt. | 1 |
How the percentage is calculated
GEO checklist score = round(passed observed weights ÷ total observed weights × 100)When every check is observed, the denominator is 15. Unknown conditions are excluded rather than automatically failed. The report also shows how many checks were observed, so a percentage should never be read without its coverage. If the page cannot be fetched safely or robots access cannot be verified, the service returns an error instead of manufacturing a score.
A score of 80 means twelve of fifteen observed weight units passed in a fully observed report. It does not mean an 80% chance of citation, 80% optimization or an 80th-percentile result. There is no competitor dataset or answer-engine measurement behind that number.
Explain a score change without inventing an outcome
Suppose a fully observed page lacks recognized authorship and a publication date, but passes the other seven checks. Those missing signals account for three weight units. Adding appropriate, truthful metadata changes the checklist, not a measured AI outcome.
Passed weights: 12 / 15
GEO checklist score: 80/100
Missing: named author (2), publication context (1)After verified metadata improvements:
Passed weights: 15 / 15
GEO checklist score: 100/100
AI citations: still not measuredWhat this GEO checker does not claim to know
The checker does not send prompts to models, inspect training datasets, query ranking systems, compare live answer citations or measure brand sentiment. It has no backlink index, competitor benchmark or Core Web Vitals measurement. It cannot establish causation between a site edit and a later mention in an answer.
Several conditions are proxies with obvious limits. A page with 149 useful words can be better than one with 151 empty words. A random external link can satisfy the reference condition. A name in metadata can satisfy authorship without demonstrating expertise. Use the evidence to ask better editorial questions instead of adding superficial fields for points.
A practical GEO improvement sequence
Verify access intentionally
Check whether the pages you want discovered are accessible under the policy you intend. Do not open restricted material merely to raise the score.
Make the main answer explicit
Use clear headings and enough server-readable explanation to answer the actual question. Add examples, definitions or evidence because they help readers, not to reach a word threshold.
Connect claims to sources and people
Name the responsible author or publisher where appropriate, cite relevant evidence and give users a way to understand the organization behind the page.
Track outcomes separately
Record what changed and when, then compare appropriate real-world evidence over time. Keep a checklist score distinct from traffic, conversion and observed citations.
Five observable signals, eight weight units
The score is the rounded share of passing observed weights. All five signals observed gives a denominator of eight. The tool does not read and judge policy quality, authenticate an author's credentials or determine whether an external source supports the claim beside its link.
| Signal | What is detected | Weight |
|---|---|---|
| Named author | Recognized author metadata or a JSON-LD author. | 2 |
| Publication context | Recognized publication-date metadata. | 1 |
| External references | At least one external HTTP(S) link. | 2 |
| About / contact references | A linked path matching the implemented about/contact pattern. | 2 |
| Editorial / privacy references | A linked path containing privacy, editorial, correction or terms. | 1 |
What experience, expertise, authority and trust still require
The appropriate evidence depends on the topic. A personal account and a technical safety instruction need different kinds of support. Treat the page's intended use as the starting point, and involve a suitably qualified reviewer where the subject warrants one.
| Human review area | Evidence to examine | Why a metadata check is insufficient |
|---|---|---|
| Experience | First-hand observations, original examples and documented methods. | An author name does not prove the work was performed. |
| Expertise | Relevant qualifications, reasoning and subject accuracy. | A biography can contain unverified claims. |
| Authoritativeness | Independent reputation and appropriate recognition. | An outbound link count is not a reputation assessment. |
| Trust | Accuracy, transparency, accountability and clear limitations. | A policy URL does not establish that a policy is adequate or followed. |
A detected author is useful, but not the conclusion
A page can have an author value and a date while leaving important claims unsupported. The analyzer should help you find the missing context, while your review checks whether the detected information is meaningful.
Detected: author, date, about link
Missing: external references, editorial/privacy reference
Passing weight: 2 + 1 + 2 = 5 of 8Checklist score: 63/100
Next review: verify the author's relevant experience, inspect factual claims and publish appropriate accountability information.Strengthen accountability rather than adding decorative badges
Explain who did the work
Where authorship matters, name the responsible person or organization and provide relevant context. Do not add credentials that cannot be verified.
Support claims at their point of use
Link to sources that directly support the statement. Distinguish your own observation from borrowed evidence and explain any uncertainty.
Use dates honestly
Show meaningful publication or revision information. Updating a date without reviewing the substance can mislead readers about how current a claim is.
Make corrections possible
Provide a credible way to contact the publisher or request a correction. A copied policy page is less useful than a process someone can actually use.
Why this is a signal analyzer, not an E-E-A-T verdict
The checker reads one server response. It does not inspect an author's history, independent reviews, professional registers, original research files or the wider reputation of a publisher. It also cannot assess whether first-hand experience is genuine or whether a claim is factually correct.
A low score can reflect a different markup pattern or a page that does not need article-style signals. A high score can be achieved with superficial metadata. Both outcomes make sense only when combined with editorial judgment and the actual purpose of the page.
AI Search Readiness Report:
questions, answered.
Are the three AI report sections based on separate crawls?
No. A single HTML fetch supplies all three sections. Robots policy is checked separately, and each heuristic exposes its evidence and limitations.
Is AI Search Readiness the same as GEO Score?
They share eight content and identity checks. GEO Score adds an OAI-SearchBot robots-access check, making its denominator 15 rather than 14 when all checks are observed.
Do I need special AI schema to pass this audit?
The tool looks for parseable JSON-LD, not a special AI vocabulary. Add accurate structured data only when it describes real visible entities.
Does the audit check answers from ChatGPT or Google AI Overviews?
No. No external AI service is called. It inspects your HTML and reports the disclosed checklist signals.
Can a page be useful for AI search with a lower score?
Yes. A page can lack a particular metadata pattern and still be useful. The checklist is a debugging and editorial aid, not a comprehensive model of retrieval quality.
Is GEO Score an official Google metric?
No. It is Outgle's disclosed weighted HTML checklist. It is not a Google score, an AI-provider score or an industry-wide standard.
Can this checker tell me whether ChatGPT cites my website?
No. It does not query answer engines or collect citation observations. A citation audit would require a separate, carefully defined sampling process.
Why does a page with strong content receive a low GEO score?
The parser may not find the specific metadata or link patterns it checks. Client-rendered content, alternative authorship markup and intentionally minimal pages can all differ from the checklist assumptions.
Should I add an author and date to every page to get 100?
Only where they are appropriate and truthful. A utility or contact page does not necessarily need article-style publication metadata. Interpret a missing signal in the context of the page's purpose.
Is this the same score Google quality raters use?
No. It is a small Outgle checklist of detectable HTML signals. It does not reproduce a quality-rater evaluation or expose an internal search-engine score.
Can the analyzer verify professional credentials?
No. A name or biography link is only a starting point. Credential verification requires appropriate independent evidence and human review.
Why does it miss the author name printed on my page?
The parser recognizes specified metadata and JSON-LD patterns. A plain visible byline or a client-rendered author component may not be detected by those rules.
Does adding a privacy link prove trustworthiness?
No. The condition detects a link path, not the policy's substance, legal adequacy or actual implementation. Read and assess the policy separately.
The calculations and checks above describe this tool's implementation. Use these primary references for the underlying formats and publishing guidance.
Google AI feature requirements and guidance ↗ (opens in a new tab)OpenAI crawler purposes ↗ (opens in a new tab)Google's people-first content and E-E-A-T guidance ↗ (opens in a new tab)Experience in the quality-rater guidelines ↗ (opens in a new tab)Put AI Search Readiness Report to work.
Fetch one public page for three disclosed heuristic reviews: AI search structure, GEO signals and visible E-E-A-T evidence.
Use the AI Search Readiness Report↑