7 patterns in how Perplexity cites Korean sites as sources

Bottom line: when Perplexity cites a Korean site as a source, seven signals keep showing up — structured data, conclusion-first copy, freshness, entity clarity, FAQ blocks, author attribution, and external mentions. A page that carries three or more of these becomes a citation candidate. The "machine-citable shape" matters more than sheer volume of content.

1. It has structured data (JSON-LD)

A page with schema.org JSON-LD — Article, FAQPage, LocalBusiness — lets AI read stated values instead of guessing facts from prose. Typed data like price, hours, and publish date is especially good at raising citation confidence. Our schema generator produces page-appropriate JSON-LD from a single URL.

2. The conclusion is at the top

When AI synthesizes an answer, it weighs the top of the page most heavily. A paragraph that opens with "Bottom line:" is itself an extractable answer fragment, so it's more likely to be quoted. Articles that bury the point under a long intro get pushed down the ranking.

3. Freshness signals are obvious

Perplexity responds strongly to "recent" queries. State the published/updated date in the body and fill the JSON-LD dateModified, and you'll be cited ahead of older pages on the same topic. A 2021 article rarely gets picked for a query seeking 2026 information.

4. The entity is clear

If it's vague who "you" are and what you do, AI hesitates to cite you. Repeating your business name, location, and category consistently across body copy and structured data sharpens the entity and raises your odds of being chosen for location+category queries like "Hapjeong hair salon Seoul."

5. There's an FAQ block

Question-answer pairs share the same structure as an AI answer, so they get quoted almost verbatim. Add FAQPage JSON-LD and the questions themselves match search queries, lifting citation odds sharply. The FAQ at the bottom of this article exists for exactly that reason.

6. The author is named

A page with a named author and a profile link is trusted more than an anonymous one. When Article JSON-LD's author links to a real person entity (@id), AI grades the page as expert-written.

7. It's mentioned elsewhere

When other sites, news, or directories mention the business or the article, its weight in AI's trust graph rises. External mentions are hard to manufacture quickly, but once the first six are in place, citations and shares follow and accumulate on their own.

Summary

You don't need all seven at once. Getting just #1, #2, and #5 (structured data, conclusion-first, FAQ) already puts you in the candidate pool. To redesign the whole structure at once, see GEO:lab consulting; to do it yourself, start with how to write llms.txt.