AEO & GEO
Structured data for search and AI answers
Schema does not buy rankings or citations. It clarifies content that users can already see. Here is what to mark up, what to leave alone, and how to validate it.
TL;DR
- Schema markup does not directly rank a page or guarantee a citation. It gives machines explicit labels for content and entities that are already visible to users.
- Start with supported types that match the page: Article for editorial content, Organization for the publisher, and BreadcrumbList for hierarchy. Use FAQPage or HowTo only when the visible content genuinely fits.
- Google states that AI Overviews and AI Mode need no special markup beyond normal search eligibility. Clear answers, authorship, evidence, and a consistent entity remain the useful work.
- Mark up what is genuinely on the page, keep it accurate, and validate it. Schema that lies or drifts from the visible content gets ignored or penalised.
- Use the llms.txt generator for a site-level AI context draft, scan agent readiness for discovery gaps, then see how we approach AEO and GEO, and our broader AEO, GEO and SEO guide for the content side.
What structured data contributes
Clear visible content first. Accurate labels second.
Search and answer systems parse headings, body copy, links, authorship, and page relationships. A direct answer near the relevant heading helps readers and also reduces ambiguity for machines. That is useful even when no rich result or AI citation appears.
Structured data labels what is already visible on the page. It labels the author, publisher, page type, hierarchy, questions, or steps when those things are present. The markup must agree with the visible page. Schema on vague, hidden, or contradictory content adds no trustworthy signal.
This is why AEO and GEO sit on top of good SEO. Google requires an indexed, snippet-eligible page for its AI features and says there are no additional technical requirements.
Mark up what is genuinely on the page, keep it accurate, and validate it. Schema that lies or drifts from the visible content gets ignored.
The schema that matters
Five types to implement before anything else.
Article
Mark up guides and posts as Article (or its subtypes). Include headline, description, author, datePublished, dateModified, and image. This tells answer engines who wrote it and when it was last reviewed, which matters for time-sensitive topics. Keep dateModified honest: update it only when you genuinely revise the content.
FAQPage
When a page genuinely answers a set of questions, FAQPage markup can describe the visible question-and-answer structure. It does not guarantee an FAQ rich result or AI citation. Only mark up questions that users can read on the page.
HowTo
For step-by-step content, HowTo markup names each step in order. Answer engines use it to reconstruct a procedure, and it pairs naturally with the way assistants present instructions. Mark up the steps that appear on the page, in the order they appear, with clear names and text.
Organization
A single, consistent Organization entity (with name, logo, URL, and sameAs links to verified profiles) helps engines connect publisher details across pages. Consistency is the point: reuse the same @id and accurate details instead of creating conflicting versions of the organization.
BreadcrumbList
Breadcrumb markup gives the engine your site structure: where this page sits and how it relates to the rest. It helps both search and answer engines understand topical context, internal linking, and page relationships. Match the breadcrumb to the visible navigation path.
At a glance
- 5
- Schema types do most of the work: Article, FAQPage, HowTo, Organization, BreadcrumbList.
- 1
- One consistent Organization entity keeps publisher details identical across the site.
- 0
- Markup that does not match the visible page earns zero benefit and risks being ignored. Never fake it.
Automation next step
Scope this as an automation: Structured data for AI answers
Tell us the manual step you want gone. We reply with what is realistic to automate and what it costs.
Implementation checklist
What to do, in order, on a real site.
✓
Lead each key page with a direct answer to the question it targets, near the top, in plain language.
✓
Add Article schema to guides and posts with author, datePublished, and an honest dateModified.
✓
Add FAQPage markup only where the questions and answers are visible on the page.
✓
Add HowTo markup to genuine step-by-step content, matching the on-page steps.
✓
Define one Organization entity with logo, URL, and sameAs, and reuse the same @id everywhere.
✓
Add BreadcrumbList that mirrors the visible navigation path.
✓
Use [Schema Generator](https://schemagenerator.dev/) to draft the JSON-LD, then validate every page with a schema testing tool and fix errors before you ship.
✓
Re-check after any redesign or CMS change, because markup breaks silently.
Common questions
What teams ask about schema and AI answers.
Does structured data directly get me cited by AI Overviews or ChatGPT?
No. Google says AI Overviews and AI Mode have no additional markup requirement beyond normal search eligibility. Schema can clarify page type, author, publisher, and visible content, but rankings and citations are selected by the search or answer system.
Which schema type should I add first?
Article and Organization. Article tells engines who wrote the page and when it was reviewed; Organization ties every page back to one consistent, trustworthy entity. After those, add FAQPage and HowTo where the page genuinely contains questions or steps, and BreadcrumbList for structure. Resist the urge to add exotic types before these five are correct and validated.
Can I add FAQ schema to any page to get more visibility?
No. Mark up only questions and answers that are actually visible on the page. FAQPage markup on hidden or invented content violates the guidelines and can get the markup ignored or the page penalised. The reliable approach is to write a genuine FAQ section that helps the reader, then mark it up. If the content is not on the page, do not mark it up.
How is this different from normal SEO?
It builds on it. The foundations are the same as good SEO: a useful direct answer, clear structure, real authorship, and trustworthy evidence. AEO and GEO put more emphasis on making answers explicit and monitoring visibility in answer formats. Schema describes the page; it does not force an answer engine to quote it. See our AEO, GEO and SEO guide for the content side.
How do I measure whether my schema is helping?
Validate first: every targeted page should pass a schema testing tool with no errors. Schema Generator can help draft the JSON-LD, but the markup still needs to match the visible page. Then track presence in AI answers and featured snippets for your target questions, and watch referral patterns from AI tools where your analytics can see them. Citations are harder to measure than rankings, so treat improvements in snippet capture and direct-answer presence as the leading indicators.
Does schema slow my page down?
No, in practice. JSON-LD is a small block of text in the page head or body and has negligible performance cost. The thing to watch is not the schema itself but keeping it accurate as the page changes, because markup that drifts from the content is worse than none. Validate after every redesign or CMS change.
How we approach it
Clear answers first, then accurate markup, then validation.
We start with the page content and add schema once the answer is clear near the top. If the page does not answer the question directly and near the top, no markup will rescue it. Once the content is clear, we add the five core schema types, keep one consistent Organization entity across the site, and validate every targeted page. Markup that matches the visible content and passes validation is the only kind that earns a benefit.
Then we keep it honest over time, because schema breaks silently after a redesign or CMS change. Read our approach to AEO and GEO for B2B teams, and the AEO, GEO and SEO guide for the content work that makes the markup worth having.
Written by
Vineet Talwar
Co-founder, Tech & Operations at Some Tech Work. WordCamp speaker across Europe and Asia, and host of the WP Shoutout podcast.
Start here
Ready to talk.Book a short diagnostic.
Tell us what needs fixing
We read every brief and reply within one business day.
Prefer to talk first?or request a tech stack audit →or email us directly →
Not sure where to start? Send the stuck decision, workflow, or page. We will say whether you need a diagnostic call, a tech stack audit, or a different first step.