What is article schema and when should I use it?

How does article schema markup actually work?

Article schema is written in JSON-LD format. A lightweight JavaScript notation that lives inside a <script type="application/ld+json"> tag in your page's HTML. It does not appear visibly on the page. Its entire audience is machines: search engine crawlers, AI bots, and any other system that reads your HTML before rendering it.

The schema block contains structured properties that describe the content. The headline, the author, the publication date, the URL, and a brief description. When AI reads your page, it doesn't have to infer these things from the copy. You've handed them over directly, in a machine-readable format it already knows how to parse.[1]

Here's a minimal, valid BlogPosting schema block for an expert node:

{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "What is article schema markup?",
  "description": "Article schema classifies your content for AI engines.",
  "url": "https://yourdomain.com/your-page",
  "datePublished": "2026-03-18",
  "dateModified": "2026-03-18",
  "author": {
    "@type": "Person",
    "name": "Your Name",
    "url": "https://yoursite.com"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Your Business Name",
    "url": "https://yoursite.com"
  }
}

This is the minimum viable schema for an content page. Adding it to every node in your directory ensures that AI systems can classify and index your content correctly. Without guessing.[2]

What types of article schema exist, and which should I use?

Schema.org defines a hierarchy of article types, each with a different signal for AI classification. Understanding the distinctions matters because using the wrong type creates ambiguity. And AI engines dislike ambiguity when they're deciding what to recommend.

Schema Type	Best For	AI Signal
Article	General editorial content, opinion pieces, research summaries	Broad. Works but not specific
BlogPosting	Expert Q&A nodes, dated knowledge posts, authority directory content	Strong. Signals dated writing
NewsArticle	Time-sensitive announcements, industry news, event coverage	Temporal. Deprioritized over time
TechArticle	Technical documentation, developer guides, API references	Specialized. Strong for tech audiences
HowTo	Step-by-step instructional content with discrete numbered steps	Procedural. Distinct from article types

For most authority directory nodes. question-based posts answering a specific query. BlogPosting is the correct choice. It signals that this is a specific, dated piece of writing, not a service page or a generic article aggregated from other sources.[3]

What properties should I include in article schema for maximum AI signal?

Not all schema properties carry equal weight. Some are required for schema to be valid; others are strongly recommended by Google's structured data guidelines; still others add incremental value. For authority nodes, here's the priority stack:

Required for validity:

• headline. Must match your H1 exactly
• author. The Person or Organization who wrote it
• datePublished. ISO 8601 format (YYYY-MM-DD)

Strongly recommended:

• description. A 1–2 sentence summary (can match your meta description)
• url. The canonical URL of this page
• dateModified. Signals freshness to AI engines
• publisher. Links the content to your organization

Additive for authority:

• image. An ImageObject with URL, width, and height
• mainEntityOfPage. A WebPage reference confirming this is the primary content
• wordCount. Signals content depth to AI classification systems

The author property deserves special attention. A bare string like "author": "Cindy Molchany" is weaker than a full Person object with a URL and sameAs links to your LinkedIn and other verified profiles. The richer the author object, the stronger the E-E-A-T signal.[4]

Why does content classification matter to AI recommendation systems?

AI recommendation engines don't just read what your content says. They read what type of content it is before they decide how to use it. This is the distinction that most experts miss entirely.

A page with no schema markup forces AI to make inferences: Is this a product page? A press release? A blog post? A service description? These inferences are slower, less reliable, and often wrong. A page with correct article schema skips the inference step entirely. AI reads the schema, classifies the content in milliseconds, and proceeds to evaluate whether your answer is substantive enough to recommend.

Schema markup is the pre-classification that earns AI's attention. Without it, your excellent content may be read. But it's less likely to be categorized correctly, which means it's less likely to be surfaced in response to the queries it answers.

The practical implication: every content page you publish should have schema on it from day one. Not added as an afterthought after the page has already been indexed without it.

How do I validate that my article schema is working correctly?

Two tools are the standard for schema validation. Use both.

Google's Rich Results Test (search.google.com/test/rich-results). Paste your page URL or raw HTML and it will show you exactly which schema types were detected, which properties were recognized, and which have errors or warnings. This is the authoritative tool for checking Google's interpretation of your markup.

Schema.org Validator (validator.schema.org). Validates your JSON-LD against the official Schema.org specification. Useful for catching structural errors that Google's tool might overlook.

What to look for when validating:

• No errors. Errors mean the schema is malformed and may not be read correctly
• No missing required fields. Warnings for missing required properties reduce the schema's reliability
• Correct type detection. Confirm the tool recognizes your BlogPosting (or whichever type you used) correctly
• Author object recognized. Confirm the Person entity with its URL and sameAs links is being parsed as a structured identity

Run validation every time you create a new node. One malformed schema block can invalidate all schema on that page.