Speakable Schema Used by Fewer Than 10% of Websites

Speakable schema delivers a 127% increase in AI citation rates but is used by fewer than 10% of websites. This is the single largest gap between impact and adoption in AI search optimisation. Speakable marks specific content sections as suitable for extraction and text-to-speech — effectively telling AI platforms "this is the answer you should cite."

What Speakable schema does

Speakable is a schema property that marks specific sections of a web page as particularly suitable for audio playback and direct extraction. Originally designed for voice assistants, it has become increasingly relevant for AI search because it explicitly identifies the most citable, extractable portions of your content.

When an AI platform encounters Speakable markup, it receives a clear signal: "The content creator considers this paragraph to be the definitive, self-contained answer." This reduces the AI's need to evaluate and select which part of the page to cite — you have already done that work for it.

The 127% citation increase

Pages with Speakable schema see a 127% increase in AI citation rates compared to equivalent pages without it. The mechanism is straightforward: Speakable removes ambiguity about which content to extract, making it easier for AI to cite your page with confidence. Easier citation equals more citations.

The 127% figure is substantial but less dramatic than FAQPage's 350% increase. However, Speakable has a unique advantage: it can be applied to any content type on any page, while FAQPage only applies to question-and-answer content. This makes Speakable more broadly applicable.

Why adoption is so low

Several factors explain the sub-10% adoption rate:

• Low awareness — most web developers and SEO professionals have never heard of Speakable
• Voice-first perception — many assume Speakable is only for voice search and smart speakers
• No Google rich result — unlike FAQPage, Speakable does not generate a visible rich snippet in Google, so there was no traditional SEO incentive to implement it
• Limited tooling — popular schema generators and WordPress plugins rarely include Speakable as an option
• AI search is new — the citation benefit has only recently been documented

How to implement Speakable

Speakable is added as a property within your Article or WebPage schema. It uses CSS selectors to point to specific content sections:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Your Article Title",
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [".answer-capsule", ".key-finding"]
  },
  "author": {
    "@type": "Person",
    "name": "Author Name"
  }
}
</script>

The cssSelector property points to the CSS classes applied to your most citable paragraphs. In the example above, any paragraph with the class answer-capsule or key-finding would be marked as speakable content.

Best practices

Practice	Why it matters
Mark 1-3 sections per page	Too many speakable sections dilute the signal
Choose self-contained paragraphs	The marked content should make sense when extracted alone
Lead with the answer	Speakable paragraphs should start with the key fact or answer
Include specific data	Numbers, percentages, and concrete claims are more citable
Combine with answer-capsule styling	Visual formatting that matches your schema markup

The competitive window

With fewer than 10% of websites using Speakable, early adopters have a clear competitive advantage. As AI search grows and more businesses discover this schema type, the advantage will narrow. The window to be among the first in your industry to implement Speakable — and capture the 127% citation lift while competitors are unaware — is now.

Read our complete schema markup guide for AI search · See all AI search statistics