How to appear in Google AI Overviews | Blog SEO Ighenatt

Google AI Overviews has changed the rules of the game in organic search. When a user makes a query, Google generates an AI-synthesised response that appears before traditional organic results, citing specific sources. The question every SEO professional asks is direct: how do I get my content to be one of those cited sources?

According to BrightEdge data, 80% of sources cited in AI Overviews already rank in the organic top 10 for that query. This means organic ranking remains the foundation, but now you need an additional layer of optimisation focused on citability by language models.

What AI Overviews are and how they work technically

AI Overviews (formerly called SGE, Search Generative Experience) is the integration of large language models (LLMs) into Google’s results. When the system detects that a query would benefit from a synthesised response, it activates the AI module that crawls, evaluates and combines information from multiple sources to generate a cohesive answer.

The technical process works in three phases. First, Google identifies candidate pages using its traditional ranking signals. Second, the AI model evaluates which fragments of those pages contain relevant, verifiable and well-structured information. Third, it synthesises a response and links the sources it has used.

In Spain, AI Overviews appear in 11% of searches since March 2025, with a growing trend that makes reaching 20-25% plausible throughout 2026. The categories with the greatest presence are health, finance, technology and education, precisely the topics where information reliability is most critical.

Factors that determine which content Google AI Overviews cites

Princeton’s research on GEO (Generative Engine Optimisation) identifies specific factors that increase the probability of an AI model citing your content:

• Verifiable quantitative data: a SEMrush study confirms that content with concrete figures attributed to sources has a 47% higher probability of being cited. It’s not enough to assert; you need data with an identifiable origin.
• Clear and demonstrable authorship: pages with an identified author, professional bio and links to verifiable profiles receive more citations than anonymous content.
• Direct answers in the first paragraphs: the AI model looks for self-contained fragments that answer the query. If your answer is buried in the fourth paragraph, it loses out to competitors who place it at the beginning.
• Clear semantic structure: H2/H3 headings that formulate questions or describe exactly the subtopic help the AI identify and extract the relevant passage.

The key difference from traditional SEO is that ranking well for your page is no longer enough. Your content needs to be citable: fragments that AI can extract and present as an answer without losing context or precision.

E-E-A-T and credibility: why sources matter more than ever

The E-E-A-T framework (Experience, Expertise, Authority, Trustworthiness) has always been relevant to Google, but with AI Overviews it becomes a much stricter filter. The AI needs to trust that the information it cites is accurate, and for that it evaluates credibility signals at multiple levels.

Demonstrable experience

Google values content created by people who have directly experienced what they describe. For an article about web speed optimisation, this means including real audit screenshots, before/after metrics from specific interventions and references to specific projects.

Documented expertise

Authorship matters. Each article should have an author with a professional bio detailing their experience on the topic. Author profiles linked to LinkedIn, academic publications or industry conferences reinforce the expertise signal.

Domain authority

Inbound links remain an authority signal, but AI Overviews particularly weights mentions and citations from recognised industry sources. An article cited by Search Engine Journal has a higher probability of being selected than one with the same content but without external mentions.

Verifiable trustworthiness

Every quantitative data point you include should have an identifiable source. Unattributed claims reduce the model’s confidence. The recommended practice is to cite the source immediately after the data point, not in a bibliography at the end of the article.

Content structure that generative AI prefers

Language models process and extract information differently from human readers. Princeton’s study on GEO identifies three formats that LLMs prefer for generating citations:

Ordered and unordered lists: when the answer to a query involves multiple elements (steps, characteristics, tools), presenting them in a list facilitates extraction. The AI can cite the complete list or select individual elements.

Comparison tables: for queries of the type “X vs Y” or “best tools for Z”, tables provide dense, structured information that AI can reformulate. A well-constructed table with clear columns and concrete data has a high probability of being cited.

Clear definitions in the first paragraph: when an H2 formulates a question (e.g., “What is Core Web Vitals?”), the immediately following paragraph should contain a concise, self-contained definition. This direct question-answer pattern is the most cited format in AI Overviews.

The underlying principle is self-containment: each section of your content should be understandable on its own, without needing to read the rest of the article. AI extracts fragments, not full pages.

Schema Markup to Increase Chances of Appearing in AI Overviews

Schema markup (structured data) does not guarantee appearance in AI Overviews, but it helps Google’s systems understand the structure and purpose of your content. Three schema types are especially relevant:

Article schema

Provides context about the editorial piece: author, publication date, editorial organisation and main topic. Google uses this data to evaluate content freshness and authorship.

FAQPage schema

Explicitly marks frequently asked question sections, helping AI identify extractable question-answer pairs. Pages with FAQPage schema show higher correlation with citations in AI Overviews according to BrightEdge analysis.

HowTo schema

For instructional content, HowTo schema structures steps so that AI can extract precise sequential instructions. Each step with its name, description and position facilitates direct citation.

Correct implementation of these schemas requires consistency between the markup and the visible content. Google penalises discrepancies: if your FAQPage schema contains questions that don’t appear visually on the page, the markup will be ignored.

Frequently asked questions about AI Overviews

Does appearing in AI Overviews increase organic traffic?

The traffic impact is mixed. Appearing cited as a source in AI Overviews can generate clicks to your website, but if the AI response satisfies the user’s query, they may not click through. Initial data shows that cited sources achieve a lower CTR than organic ranking #1 but higher than positions 5-10.

Can I control whether my content appears in AI Overviews?

You cannot choose to appear, but you can opt out using the nosnippet meta tag or max-snippet:0. To maximise the probability of appearing, optimise your content following E-E-A-T guidelines and structure the information in a clear and citable way.

Do AI Overviews replace featured snippets?

They don’t replace them, but they coexist. In some queries Google shows AI Overviews instead of featured snippets, in others it shows both. The trend is for AI Overviews to expand to more query types, but featured snippets continue to exist for specific searches.

Optimising for AI Overviews requires a technical approach that combines E-E-A-T, structured data and content designed to be cited by AI models. If you need your website to appear in Google’s AI-generated responses and want a strategy tailored to your sector, contact our team for a GEO citability audit.