A little, indirectly, for reasons neither Google nor Microsoft has fully spelled out. It's not the reason AI engines skip you, and it's not the reason they'd pick you either.
Schema.org is a shared vocabulary. You wrap facts about your page in JSON-LD - a script tag full of structured key-value pairs - so a machine doesn't have to guess what a paragraph means. "Organization" markup states your company name, logo, and address as data, not sentences. "Product" states price and availability. "FAQPage" states which sentence is the question and which is the answer. "Article" states author, publish date, and headline.
None of that is new. It's the same vocabulary that's powered Google's rich results - star ratings, FAQ dropdowns, recipe cards - for over a decade. What's changed is the audience reading it. It used to be Googlebot deciding whether to show a rich snippet. Now it's also a retrieval pipeline deciding what to hand an AI model as source material for an answer.
Google's own AI features documentation is direct about this: "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add." There are no additional technical requirements for AI Overviews or AI Mode beyond standard Search eligibility.
Google's structured data intro doc explains the broader purpose: it says structured data helps Google "understand the content of the page, as well as to gather information about the web and the world in general, such as information about the people, books, or companies that are included in the markup." That's a real, documented mechanism - it's just described for search and knowledge understanding generally, not as a specific AI-citation lever.
One concrete data point that did move: Google retired FAQ rich results from its search results pages on May 7, 2026, after years of restricting the feature to authoritative sites. The FAQPage schema type itself is still valid and won't break anything if you keep it - the SERP dropdown is just gone. For AEO purposes that dropdown was never the point; the value is the question-answer structure itself, which is the shape a retrieval step is looking for regardless of whether Google renders it in blue-link results.
Microsoft's own guidance is closer to an endorsement than Google's, but it still stops short of "ranking factor." Bing's crawler reads Schema.org markup - Article, FAQPage, HowTo, Organization, Person - and uses it to classify content type, author, and topical scope. That's consistent with what we covered in the Bing Webmaster Tools AI Performance report: Bing already reports which of your pages get read as grounding sources for Copilot answers, which means its systems are actively parsing page content, structured or not, to build those answers.
Schema doesn't make Copilot cite you. What a connected entity graph does - Organization and Person tied together, Article on the post, FAQPage on the actual questions - is remove ambiguity about what entity is making a claim. That's a clarity improvement, not a visibility guarantee.
Here's the part that's genuinely useful and rarely stated plainly: AI retrieval pipelines don't read your page the way a human does. They parse raw HTML, often at speed, often stripping formatting, sometimes truncating content before your best sentence. When a fact is stated only in prose - "we're based in Dubai and our starter plan runs 450 AED a month" - a parser has to infer structure that isn't there. When the same fact is also present in JSON-LD as explicit fields, there's nothing to infer. It's already labeled.
That's the actual case for schema in AEO: it reduces the odds your entity facts get misread, misattributed, or dropped entirely during automated parsing - especially facts that matter for buyer questions like who/where/price/rating. It's a hedge against machine error, not a ranking input. No engine has published a study saying "pages with schema get cited X% more" that would survive scrutiny of its methodology.
Schema genuinely helps on answer-shaped pages: FAQ sections that already contain real question-answer pairs, product pages with real prices and availability, Organization markup that keeps your name, address, and category consistent across every page you own. It's cargo cult when it's bolted onto thin or vague content in the hope that labeling weak prose will make it citable. It won't. A retrieval pipeline that finds a well-formed FAQPage schema wrapped around a vague, evasive answer still has a vague, evasive answer to work with.
We don't ship schema as a deliverable and call it AI visibility. We check whether a site is structurally readable at all - see the AI readiness checklist for what that covers - then we measure the thing schema can't tell you: whether a frozen set of real buyer questions, asked across ChatGPT, Claude, Gemini, and Perplexity, actually names your brand. Clean markup reduces the odds you're misread. It doesn't tell you whether you're being picked. Those are two different problems, and only one of them shows up in a citation count.