How to format your blog post for Perplexity AI citations
Perplexity AI cites sources that answer questions directly and concisely. Learn the exact formatting signals — structure, prose style, and metadata — that put your posts in its answer cards.
Perplexity AI cites your blog post when it detects a direct, factual answer near the top of the page — ideally in the first two sentences after the heading. The algorithm is looking for specificity and confidence, not hedging or preamble. If your opening paragraph starts with 'Great question — in this article we will explore...', Perplexity will skip you.
Why does Perplexity choose one source over another?
Perplexity's retrieval model works like a very fast reader skimming the page for a usable quote. It wants a sentence that (a) directly answers the query, (b) is factually specific — numbers, dates, concrete steps — and (c) appears early enough that the retriever doesn't have to read the whole article. Blog posts that bury the answer after two paragraphs of context-setting lose to posts that lead with the answer, every time.
The most reliable signal is what AEO practitioners call the 'inverted pyramid' — borrowed from journalism. Put the most important fact first, then expand. If a reader (or a model) reads only your first sentence, they should still get a usable answer. Everything after the first sentence is supporting detail.
The direct-answer paragraph pattern
Here's the pattern that gets cited. Open each major section with one sentence that answers the section heading as if it were a question. Keep it to 20–35 words. Then elaborate for the human reader. This dual-audience writing — precise enough for a language model, rich enough for a person — is the core discipline of AEO.
In VeloCMS, the SEO score panel grades your 'first sentence answer quality' live as you type. Watch for the AEO indicator in the bottom-right of the editor — green means your opening paragraph is citation-ready.
Metadata signals Perplexity reads
Beyond body copy, Perplexity reads the page's JSON-LD Article schema — specifically the headline, description, and dateModified fields. A fresh dateModified signals recency, which Perplexity weights heavily for news-adjacent queries. VeloCMS automatically sets dateModified to the last publish time, so you don't need to manage this manually. What you do need to manage is keeping your meta description in sync with the direct-answer paragraph — when they match, Perplexity's confidence in the citation goes up.
What to avoid
Avoid framing every section as a teaser rather than an answer — 'In this section we'll cover...' is a teaser. 'The fastest way to X is Y because Z' is an answer. Avoid jargon-heavy intros that require a domain expert to decode. And avoid passive voice for key claims: 'It is generally understood that...' gets skipped; 'Cache-Control: max-age=31536000 tells browsers to hold static assets for one year' gets cited.
Related: see 'How AI search engines crawl Next.js sites' for the technical side — RSC rendering, meta tags, and sitemap freshness all feed into whether Perplexity can even reach your content.