Google AI Overviews appear at the top of search results for a growing share of queries, above every ranked organic result. If your content isn’t being sourced in them, you’re competing for clicks after buyers have already read an AI-generated answer that didn’t include you.

The signals that determine inclusion are specific and improvable. AI Overviews selection isn’t purely a function of domain authority or existing rankings. A page with strong FAQPage schema and clear, direct-answer formatting can contribute to an AI Overview even without ranking in position one. And pages ranking in the top three get skipped regularly when they’re not structured in a way Google’s system can extract.

This post covers the four actions most likely to improve your AI Overviews inclusion, in the order that typically produces results fastest. If you’re newer to the broader discipline of optimizing for AI-generated answers, the complete guide to generative engine optimization covers the full framework this post builds on.

What Actually Determines AI Overviews Inclusion

Google hasn’t published a full technical specification for the AI Overviews selection. What’s observable from Google’s structured data documentation and patterns across large numbers of pages is that four things matter consistently:

Schema markup. Specifically, FAQPage and Organization JSON-LD. These give Google’s system explicitly labeled, machine-readable content to work with rather than requiring inference from unstructured HTML.

Entity graph completeness. Google’s Knowledge Graph connects entities (companies, products, topics) to verified facts. Brands with complete entity graph entries (Google Business Profile, Wikipedia/Wikidata where applicable, Organization schema with sameAs links) are described and sourced with more confidence.

Content structure. Pages that answer questions directly at the start of each section (before supporting context) are easier for AI systems to extract from. The underlying mechanism is that AI Overviews source from content chunks, and a chunk that leads with the answer scores higher for relevance than one that buries it.

Topical authority signals. Pages on sites with deep, internally linked coverage of a topic are more likely to be sourced. This isn’t just domain authority in the traditional SEO sense. It’s about whether Google’s system recognizes your site as genuinely covering the topic space the query belongs to.

Traditional search ranking still matters as a baseline. Pages Google hasn’t indexed or doesn’t consider relevant won’t appear in AI Overviews. But ranking alone doesn’t get you in, and failing to rank in position one doesn’t keep you out.

Step 1: Implement FAQPage Schema on Your FAQ Content

The single highest-leverage technical action for AI Overviews is adding FAQPage JSON-LD schema to pages that already contain question-and-answer content.

Google’s structured data documentation states that FAQPage schema may help with content discovery and eligibility for rich results. It doesn’t guarantee AI Overviews inclusion, but across the audits we run, pages with correct FAQPage schema show up in AI Overviews at a meaningfully higher rate than the same content without it. The schema gives Google’s system an explicitly labeled Q&A pair to evaluate rather than requiring it to find and extract a question-answer structure from prose.

Start by checking whether your FAQ pages and FAQ sections already have FAQPage schema implemented. Run your URLs through Google’s Rich Results Test. If FAQPage schema isn’t detected, it either isn’t implemented or has errors.

If it isn’t implemented, add FAQPage JSON-LD to the page <head>. The schema should list each question and its direct answer as a Question entity with an acceptedAnswer. The answers in the schema must match the visible answers on the page. Google validates against the rendered content, so a mismatch between your schema and your HTML will get flagged.

The most common mistake we see: teams implement FAQPage schema correctly but leave the actual FAQ content buried at the bottom of a long page. The schema tells Google the Q&A pairs exist. The page structure determines whether they’re actually findable. If a crawler has to read through 2,000 words of narrative to find the visual match for your schema, it’s going to have less confidence in the signal. Both the markup and the content placement need to work together.

Step 2: Audit and Complete Your Google Entity Graph

Google AI Overviews draws on Google’s own Knowledge Graph, not just page content. If your brand isn’t well-represented in that graph, the system has less information to work with when generating answers that should include you.

Three components of the Google entity graph are worth checking:

Google Business Profile. If you haven’t claimed and completed yours, do it now. Set the primary category accurately, add your website URL, and verify the listing. This is the most direct signal to Google’s system that your business exists and is categorized correctly.

Organization schema with sameAs links. The sameAs property in your homepage’s Organization JSON-LD block declares the connection between your website and your external profiles: Wikipedia, Wikidata, Crunchbase, LinkedIn, and G2. This is how you explicitly tell Google’s entity resolution system that all of these profiles refer to the same entity. Without it, Google has to infer these connections, and inferred connections are weaker signals.

Wikipedia and Wikidata. Wikipedia is heavily weighted in how Google’s Knowledge Graph is built. If your brand has a Wikipedia entry, Google has a structured source to draw from. Wikidata is the machine-readable companion. It allows AI systems to query specific facts about your entity directly. Not every company qualifies for a Wikipedia article (notability requirements apply), but Wikidata entries have a lower bar and are worth creating even when Wikipedia isn’t yet accessible.

The practical check: search your brand name on Google and look at the Knowledge Panel on the right side of the results page. If no panel appears, or if the panel has sparse or incorrect information, your entity graph has gaps.

This is where most teams have the fastest improvement opportunity with the least amount of new work.

Step 3: Restructure Your Highest-Traffic Pages for Direct-Answer Format

AI Overviews source from content that answers questions clearly and immediately. If your existing pages cover the right topics but bury the answers deep in paragraphs, they’re harder for Google’s system to extract and use, even if they rank well and get significant organic traffic.

The fix is structural, not a content rewrite. Open your highest-traffic informational pages and look at the first sentence under each H2 heading. If that sentence doesn’t directly answer the question implied by the heading, restructure it so it does. Move the answer up. Everything that was providing context before the answer can come after it.

Here’s what that difference looks like in practice. A heading that says “How We Handle Security” followed by “Our team takes a multi-layered approach to security across all customer environments…” is set up for a preamble. An AI Overview needs something more like “All customer data is encrypted at rest and in transit using AES-256 encryption.” The second version is extractable. The first requires the system to keep reading to find the actual answer. AI Overviews don’t keep reading.

This applies to every section on every page that’s a potential AI Overviews source. It takes editing time, not content creation time. For most sites, there’s significant room to improve before any new pages need to be written.

Step 4: Build Topical Authority on Queries That Trigger Overviews

Google AI Overviews appear most commonly on informational queries: “what is,” “how to,” “what are the best,” comparison queries, and definition queries. Transactional queries and highly navigational queries trigger them less often.

If your site has thin or disconnected coverage of the topic cluster your target queries belong to, Google’s system is less likely to source from you even on queries where your content is directly relevant. This is related to the traditional SEO concept of topical authority, but applied specifically to the content types that AI Overviews draw from.

The practical version: map the informational queries your buyers ask and check whether your site has clear, interlinked content covering that topic cluster. If your site has a single page on a topic and competitors have five deeply interlinked pages covering the same topic from different angles, they have a topical authority advantage that schema and content structure alone won’t overcome.

This step takes the most time of the four because it involves content creation. That’s why it comes last. Schema fixes, entity graph completion, and content restructuring are all faster and often sufficient to see movement in AI Overviews. Build topical depth as a second phase once those wins are in place.

How to Track Your AI Overviews Performance

Two tracking methods work in parallel:

Google Search Console. The Search Appearance filter shows which pages have FAQPage rich result eligibility. A page appearing here means Google is reading your FAQPage schema. It’s a leading indicator that the page is a candidate for AI Overviews sourcing, though it doesn’t confirm inclusion.

Manual query testing. Once a month, run your target informational queries in Google (incognito browser, clean session) and record whether an AI Overview appears, whether your brand or content is sourced in it, and which competitors appear. This is the direct measurement. It’s the only way to confirm actual inclusion rather than eligibility.

Changes to the schema tend to show up in Search Console within one to two weeks. Changes to AI Overviews inclusion take longer to observe because they depend on Google’s crawl cycle and how frequently AI Overviews are refreshed for specific queries. Testing monthly gives you enough data to see whether changes are having an effect without over-indexing on session-level noise.