Schema markup is a small block of code added to a page that describes what the content means in a format machines can read directly, instead of inferring meaning from paragraphs and headings the way a human does. It doesn't change what a visitor sees. It sits alongside the visible page and tells crawlers things like "this is a recipe," "this is a review with a 4.5-star rating," or "this is a blog post published on this date by this author."

Every post on this site carries schema markup identifying it as a BlogPosting, with the headline, author, and publish date spelled out explicitly rather than left for a crawler to guess from the page layout.

Where it started, and why it's changed

For years, schema markup's main job was earning rich results: star ratings under a review, a recipe card with cook time, an FAQ dropdown right in the search listing. In August 2023, Google retired the FAQ and How-to rich result types, and sites that had leaned on schema purely for that visual payoff lost the reason they'd added it.

What replaced that reasoning is bigger. In March 2025, both Google and Microsoft publicly confirmed they use schema markup to power generative AI search features, not just traditional rich results. Google's own guidance is direct about it: structured data is valuable because it's precise and easy for machines to process, which matters more, not less, as more of search happens through AI-generated answers instead of a list of links.

What this looks like in practice

An AI-generated answer has to decide, in the moment, which page to pull information from and how to represent it. Schema markup gives it a shortcut: instead of parsing prose to figure out an author, a price, or a publish date, the AI system reads a structured field that states it directly. Google's May 2025 guidance for AI search specifically calls out making sure structured data matches the visible content on the page, since a mismatch between what the schema claims and what's on the page is treated as a trust problem, not just an SEO oversight.

This is also part of why Google Search Console now tracks AI Mode and AI Overview performance using the same reporting Google already uses for standard search results. If schema markup helps determine whether your page gets pulled into an AI answer, watching that traffic in Search Console is how you'd know it's working.

Common schema types worth knowing

  • Article / BlogPosting. Identifies headline, author, and publish date for written content.
  • Organization. Establishes your business name, logo, and official properties, useful for AI systems trying to confirm who's publishing the content.
  • FAQPage. Marks up question-and-answer content directly.
  • Product. Covers price, availability, and reviews for ecommerce pages.
  • LocalBusiness. Covers hours, address, and service area for a physical or local business.

You don't need every type. Match the schema to what the page is, and skip markup for content types you don't have.

The short version

Schema markup used to be an SEO tactic aimed at rich results. It's now closer to infrastructure: a way of stating plainly, in a format AI systems can read without guessing, what a page is and what it means. Google and Microsoft have both said as much directly. Adding accurate schema costs nothing but a bit of setup time, and it's one of the more durable technical SEO habits worth building.

Want the rest of the AI-visibility picture? Read what an llms.txt file does, or check how Google Search Console now tracks AI Overview performance directly.

Sources