Five layers, one evidence-led audit
Seryvon groups website audits into five pillars: SEO, GEO, GSO, AEO and ASO. Each addresses a different discovery or usage question. Together they help teams separate technical accessibility, content readiness and measured external visibility before deciding what to change.
Seryvon is a public open-core release. The managed product is coming soon. These descriptions explain the current source implementation, not a promise of hosted availability or search results.
Choose the question you need to answer
| Pillar | Question | Example evidence |
|---|---|---|
| SEO | Can a crawler access and interpret the page? | Titles, canonical tags, headings and crawl links. |
| GEO | Is content retrievable and is it cited in tested responses? | Raw HTML parity and configured citation measurements. |
| GSO | What signals support review of Google AI search visibility? | Structured content and provider-supplied AI Overview observations. |
| AEO | Can a reader or engine extract a clear answer? | Definitions, lead paragraphs, attribution and tables. |
| ASO | Can an agent discover interfaces and understand actions? | Form labels, action markup and API documentation links. |
How to use the pillars together
Consider a documentation page that looks complete in a browser but serves very little text before JavaScript runs. Start with SEO to review access and page structure. Use GEO to inspect the raw-versus-rendered content evidence. Then review AEO to see whether the instructions begin with a direct answer and state their prerequisites. These are related checks on the same page, not five independent explanations for its traffic.
For an appointment website, add ASO when the task involves a booking form. A labelled form and declared reservation action provide static clues; a successful booking still needs a separate, authorized interaction test. GSO becomes useful when the investigation specifically concerns Google AI search observations. It should not turn an ordinary content review into a claim that the site appears in an AI Overview.
Read coverage alongside every score
The current scoring engine calculates a weighted mean of measured criteria and multiplies the pillar result by measurement coverage. Criteria marked not applicable are excluded from the eligible base. Missing measurements therefore differ from measured failures, although reduced coverage can lower the displayed score.
Some criteria contribute to several pillars. Structured data, for example, appears in SEO, GSO, AEO and ASO. Do not add pillar scores together or interpret a shared signal as several independent wins. Inspect the underlying criterion and affected pages before creating duplicate work.
Compare like with like
The implemented comparison module distinguishes exact, compatible, intersection and incompatible measurement profiles. Changing thresholds or providers can change what a score means. Preserve the crawl scope and configuration when evaluating a fix, and review coverage changes separately from content changes.
Start with the quickstart, then read the methodology and limitations. Readiness is an audit interpretation; ranking, citations and completed agent tasks require their own observations.