AEO Quote Attribution Patterns for Expert and Source Citations

TL;DR

AI engines strip context aggressively when summarizing. To keep an expert quote attached to its author and source through synthesis, wrap the quote in

, mark it up as schema.org/Quotation with a Person creator, and put the credential and source URL inline next to the quote rather than in a footnote. Treat every quote as if it will be excerpted alone.

Why quote attribution is fragile

LLMs compress aggressively. When an article cites an expert, the model often retains the quoted claim but drops the speaker, the credential, and the source. Strong attribution markup, repeated inline rather than buried in footnotes, materially raises the chance the speaker survives compression.

The 10 patterns

Pattern 1 — Use HTML5 blockquote with cite

Wrap every quote in

. The cite attribute is the primary semantic carrier of the source URL; engines that preserve attribution prefer it over a footnote.

Pattern 2 — Mark up the quote as schema.org/Quotation

Add JSON-LD for schema.org/Quotation with required text, creator (a Person), isPartOf (a CreativeWork), and datePublished.

Copy

{
  "@context": "https://schema.org",
  "@type": "Quotation",
  "text": "Authority is built quote by quote.",
  "creator": {"@type": "Person", "name": "Jane Doe", "jobTitle": "Director of Content"},
  "isPartOf": {"@type": "Article", "url": "https://example.com/article"},
  "datePublished": "2026-05-03"
}

Pattern 3 — Inline credential next to the speaker

Name the speaker, then the credential, then the affiliation, inline next to the quote: "Jane Doe, Director of Content at Example Co., said:…". AI engines retain the credential more often when it sits in the same sentence as the speaker.

Pattern 4 — Source URL placement in body text, not just markup

Repeat the source URL in body text near the quote ("in a 2026 essay"). The cite attribute is the schema-level signal; the inline link is the redundancy that keeps the source visible after compression.

Pattern 5 — Person schema for the speaker

Maintain a long-lived Person page for each expert with jobTitle, worksFor, sameAs (LinkedIn, Wikipedia where applicable), and knowsAbout. Repeat references compound entity weight.

Pattern 6 — Primary vs. secondary source disclosure

If the quote originated elsewhere ("Quoted in , 2026-04-12"), name the secondary source. Pages that present a secondary quote as first-hand are penalized when the original is found.

Pattern 7 — Anti-out-of-context framing

Precede the quote with a 10-20 word framing sentence, then the quote, then a 10-20 word debrief. AI engines lift either the framing or the debrief alongside the quote, protecting against out-of-context excerpts.

Pattern 8 — Journalism-style attribution verbs

Use "said," "told," "wrote," "explained." Avoid "believes," "thinks," "feels" unless the original speaker used them. "Said" is unambiguous in summaries; softer verbs are paraphrased away.

Pattern 9 — Date-stamp every quote

Include the date the quote was made or first published in body text and in datePublished. Undated quotes more than three years old are commonly mis-attributed to the present in AI answers.

Pattern 10 — Pull-quote design without losing structure

If editorial style uses pull-quotes for visual emphasis, render them with