Here, “AI verification” means making clear sourcing, evidence of expertise and author background available for readers to inspect; it does not mean that an AI has verified the author’s expertise.
To structure expertise for AI verification, connect substantive claims to clear sourcing and identifiable expertise, with links to relevant author or company background (Creating helpful, reliable, people-first content). You can then make the work behind a testing claim inspectable through its method, results and supporting material. Connect that visible account to structured data that matches what readers see on the page (AI features and your website).
Start with the groundwork in our earlier article: AI citations: make claims easy to verify
How do I structure an expert-led page so readers and AI systems can trace its claims to people and evidence?
Keep important content available as text, rather than leaving the substantive explanation entirely inside images or video (Google, AI features).
For background about the people or organization responsible, link to a relevant author page or the site's About page (Google, Creating helpful). These links should serve the reader's assessment of trust, not substitute for sourcing the information presented in the body.
When reporting product tests, explain what was tested, the results and how the tests were conducted, alongside evidence of the work involved (Google, Creating helpful). Google's Creating helpful, reliable, people-first content gives photographs as an example of evidence that can accompany that explanation. Use that distinction to separate the account of the test from the material showing the work: readers should have both, not a photograph standing in for the method.
Put a byline on a page where readers would expect to see one (Google, Creating helpful). Make that byline lead to background about the author and the areas they write about, rather than treating the name alone as the full account of expertise. For a product review, state how many products were tested so readers can understand the extent of the testing behind the reported results. Keep that count alongside the testing explanation, rather than asking readers to infer the scope from the products pictured.
Make the expert-led page easy to find through internal links elsewhere on your website. Where images or videos support the explanation, choose high-quality material appropriate to the content.
Expert-led page structure: implementation and checks
| Step | Documented basis | How to check the page |
|---|---|---|
| Connect claims to sources and relevant background | Google asks for clear sourcing, evidence of expertise and author or site background. | Check that readers can follow sourcing and links to relevant author or About information. |
| Expose the work behind a test-based claim | For product reviews, Google describes test counts, results, methods and evidence such as photographs. | For a test-based review, check that the number tested, findings, procedure and supporting evidence are visible. |
| Display a byline where expected | Google asks whether pages carry a byline where readers would expect one. | Check that an article's creator is named visibly. |
| Link the byline to author background | Google asks whether bylines lead to background about authors and their subject areas. | Follow the byline link and check that the destination explains the author's background and areas of writing. |
| Keep essential information in text and markup consistent | Google recommends textual availability of important content and structured data that matches visible text. | Check that important information is available as text and compare markup with the visible page. |
How should I connect the visible authors and publisher to structured data?
Make the structured data agree with the identities presented on the page (Google, AI features). Include everyone visibly presented as an author in the author markup, rather than marking up only a selected contributor (Article (Article, NewsArticle, BlogPosting) structured data).
When specifying multiple authors in markup, list each author in their own author field (Google, Article structured data).
Use Person for a human author and Organization for an organizational identity (Google, Article structured data). Where the page distinguishes an individual writer from an organizational publisher, preserve that distinction through the appropriate types. Do not represent a company publisher as a person merely to make the page appear individually authored.
Give author.url a destination that uniquely identifies the author of the article (Google, Article structured data). Google also understands sameAs as an alternative to url when disambiguating authors. Its author-markup guidance strongly recommends type together with url or sameAs to help it understand who the author is.
Use author markup to help Google understand who the author is; do not treat that identity markup as certification of credentials or proof of an individual claim.
Do not manufacture the person behind the byline through invented names, false credentials or AI-generated headshots used to simulate human experts (Google, Creating helpful). Google's helpful-content guidance describes fabricated creator profiles of that kind as deceptive and says deception makes a page untrustworthy to users and its automated quality systems. Treat deceptive authorship as a page-quality issue, not merely an incomplete profile: Google describes deception as a signal of a low-quality page.
What can this structure actually establish about AI discovery, verification and citation?
Treat the structure as a way to make expertise inspectable, not as a promise that Google will select the page (Google, AI features). Google's AI Search guidance does not guarantee crawling, indexing, serving or citation, even when a page meets requirements and best practices.
Google's AI features guidance says AI Overviews and AI Mode may use query fan-out, issuing searches across subtopics and data sources while developing a response. During generation, the models identify additional supporting pages, allowing a wider and more diverse set of helpful links than classic web search.
Do not expect AI Mode and AI Overviews to return identical responses or links (Google, AI features). They may use different models and techniques, and AI Overviews appear only when Google's systems determine they add to classic Search.
Structured data also has a separate limit: Google does not guarantee that search features consuming it will appear (Google, Article structured data). E-E-A-T itself is not a specific ranking factor, so do not turn the presence of author information into a claim that you have added one (Google, Creating helpful). Its systems give more weight to content aligned with strong E-E-A-T when topics could significantly affect people's health, financial stability or safety, or society's welfare or well-being. Use those consequences to assess whether a B2B page covers a Your Money or Your Life (YMYL) topic.
Sources
Creating helpful, reliable, people-first content, Google, 5 October 2026developers.google.com
AI features and your website, Google Search Central, 10 December 2025developers.google.com
Article (Article, NewsArticle, BlogPosting) structured data, Google Search Central, last updated 8 September 2026developers.google.com
