GSO: visibility in Google AI search experiences
Seryvon uses GSO to organize checks relevant to reviewing Google AI search visibility. The current implementation combines structured-content signals, available performance inputs and provider-supplied AI Overview metrics. It does not turn a markup check into proof that a page is eligible for, or displayed in, a particular Google experience.
Scope
The GSO scoring source checks for FAQPage, HowTo and BreadcrumbList types. It also detects an ItemList or HTML table and question-and-answer formatting. These are mostly site-level presence checks: finding the relevant structure on one crawled page can satisfy a criterion. They do not certify consistent implementation across every page.
Use those findings to review whether the content structure serves the page. A comparison table is useful when readers need to compare the same attributes across options. A procedural page needs actual steps and prerequisites. Do not add unrelated schema types simply because their presence changes an audit score.
What does a performance result mean?
The criterion named cwv_eligible requires all three supplied metrics: LCP, CLS and INP. The source compares them with its thresholds of 2,500 milliseconds, 0.1 and 200 milliseconds respectively. If required measurements are missing, the result is not measured. The criterion name describes an internal rule; it is not a Google-issued eligibility decision.
When investigating slow content, record where the supplied performance data came from and which URL or scope it represents. Check the page under the relevant conditions before attributing a change to a specific fix. A performance threshold and an AI search observation are separate pieces of evidence.
Separate observations from readiness
The AI Overview rule reads presence rate, trigger rate, average position, query count and provider from available external metrics. If those metrics are absent, it returns not measured. The current long-tail coverage rule also returns not measured; its existence in the catalogue should not be presented as completed keyword-coverage analysis.
An observed AI Overview presence rate applies to the collected query set and provider output. It is not the probability that every searcher will see the website. Nor does an AI Overview observation establish visibility across every AI Mode interaction. Keep the original query context and collection date with any interpretation.
A concrete review sequence
- Choose a real informational page and confirm its explanation is available to the crawler.
- Inspect its visible structure and any detected markup. Verify that lists, questions and breadcrumbs describe real content.
- Review available performance evidence and record missing inputs explicitly.
- If provider observations are available, inspect the queries and cited destinations separately from the structural score.
- Revisit the same scope after an editorial or technical change. Avoid attributing a visibility change to that change alone without supporting evidence.
The GEO guide explains broader citation evidence, while AEO focuses on extractable answers. See the Example report and Quickstart for the audit workflow.
Seryvon is a public open-core release published by Powehi. The managed product is coming soon. Scores do not guarantee rankings, citations or inclusion in Google AI search.