Your Schema Markup Matters More Than Your Copywriting Now.

I say this as someone who genuinely cares about good writing, and who’s spent a lot of this blog arguing that specific, well-argued opinions get cited over generic content. All of that’s still true. But there’s a less glamorous factor sitting underneath it that I think matters even more right now, and almost nobody’s talking about it at dinner parties: whether a machine can actually parse your page’s structure cleanly enough to extract the answer in the first place.

You can write the most beautifully specific, well-reasoned paragraph on the internet. If it’s buried in unstructured HTML with no schema markup, unclear headers, and no machine-readable signal about what the content actually is, an AI model may simply fail to extract it cleanly, and cite a worse-written competitor whose page happens to be structured properly instead.

Why This Sounds Backwards

Great writing assumes a human reader. AI extraction assumes structure.

Good prose is built for humans, who bring context, skim intelligently, and infer meaning from tone and flow. AI extraction systems don’t read the way humans do. They parse. They look for structural signals, headers, schema markup, clearly delimited answer blocks, FAQ formatting, explicit product or article metadata, to reliably identify what a piece of content actually is and pull the relevant part out cleanly.

A page with excellent writing and zero structural signal is genuinely harder for these systems to extract cleanly than a page with mediocre writing wrapped in clean, well-implemented schema. That’s an uncomfortable thing to say to a room full of copywriters, myself included on my more idealistic days, but I’ve watched it play out enough times to believe it.

The Actual Opinion

Technical structure is not the boring prerequisite anymore. It’s the deciding factor.

For years, schema markup and structured data were treated as a nice-to-have technical SEO checklist item, something the dev team handled once and forgot about. I think that’s now backwards. In a world where AI extraction determines who gets cited, structured data is arguably more directly tied to visibility than the quality of the prose sitting inside it. Not more important than having something good to say. More important than most teams currently treat it.

What To Actually Fix

Practical structural priorities

• Implement Article, FAQ, Product, and Organization schema markup wherever genuinely relevant, not as a checkbox exercise but accurately reflecting the actual content.
• Use clear, hierarchical headers (H2s and H3s that actually describe the content beneath them) instead of vague or clever section titles.
• Build explicit FAQ sections with direct question-and-answer formatting for the questions your audience actually asks.
• Audit your most important pages by viewing them stripped of design, just structure and text, and ask honestly whether a machine could cleanly extract “the answer” from what’s left.

Quick Answers

A few direct questions, answered directly

What is schema markup and why does it matter?
Schema markup is structured data added to a webpage that helps search engines and AI systems understand and accurately extract its content. It’s increasingly important for being cited in AI-generated answers.

Does content quality still matter for AI search?
Yes, quality and specificity still matter significantly. The point is that structural clarity is what determines whether that quality can actually be extracted and cited in the first place.

How do I make my content easier for AI to cite?
Implement relevant schema markup, use clear hierarchical headers, structure direct answers in FAQ format, and ensure content is organized in a way that’s easy to parse independent of visual design.