Does the intent-first framework work for existing articles?

Yes — it is most valuable as a retrofit. Move the direct answer into the first paragraph, add a TL;DR, convert headings to questions, and prune off-intent FAQs. Restructured pages typically show movement in Perplexity within weeks.

How long should the direct answer at the top be?

Two to four sentences, 40–100 words, written like a dictionary definition. It must resolve the query on its own — supporting context belongs below it, never above it.

Should every section use FAQPage schema?

No. Mark up only the FAQ section, and only when the questions and answers are visible on the page. Schema mirrors content; it does not substitute for it.

Is the TL;DR redundant with the direct answer?

They serve different readers: the answer resolves the query in prose an extractor can quote; the TL;DR gives skimmers the two or three takeaways beyond the core answer. Keep them non-identical or cut the weaker one.

The Intent-First Framework: One Article Structure for SEO, AEO, and GEO

The intent-first framework is a five-part article structure — direct answer, TL;DR, in-depth explanation, intent-matched FAQ, and edge cases — ordered by how much of the reader's question each part resolves. It exists because the three disciplines competing for your attention reward the same thing at different depths: search engines rank the depth, answer engines extract the opening, and AI engines cite the nuance. One structure serves all three; you write the article once.

TL;DR

Open with a 2–4 sentence dictionary-style answer — direct declarative answers lifted AI citation rates 109% in GeoSource's 540-check study.
Order the rest by reader commitment: 3-line TL;DR for skimmers, question-headed depth for researchers, intent-matched FAQ for follow-ups, edge cases for experts.
The FAQ intent test: would someone who searched this exact query type this follow-up next? If not, cut the question.

Why order sections by intent instead of narrative?

Because nobody consumes an article linearly anymore. A featured snippet quotes one paragraph. An AI answer extracts one passage — and about 85% of pages AI engines retrieve never make the final answer, mostly because the extractable answer is buried. A skimmer reads three bullets. A researcher reads everything. Narrative structure ("let me set the context first") optimizes for a reader who no longer exists; intent structure hands each real reader their layer immediately. We covered the retrieval mechanics behind this in GEO vs SEO — the short version: rankings get you retrieved, structure gets you quoted.

What are the five sections?

#	Section	Reader	Engine signal
1	Answer the intent (2–4 sentences)	Everyone	AEO: the extracted "quick answer"; +109% citation lift for declarative openings (GeoSource)
2	TL;DR (3 lines)	Skimmers — the majority	Social: skimmable summaries are what travel on Reddit, X, LinkedIn
3	In depth (question-phrased H2s, sourced stats, tables)	Researchers	SEO: long-form structure ranks for tail queries; question headings match the query fan-out
4	FAQ (intent-matched only)	The almost-satisfied	AEO: answer-box eligibility via visible Q&As
5	Misc / edge cases	Experts, unusual situations	GEO: nuanced caveats are the "non-commodity content" Google's AI guidance says earns citations

How does each section earn its keep?

Section 1 is the hardest discipline: answer the exact question, dictionary-style, before any context. Section 3 carries the evidence — the Princeton GEO research (KDD 2024) measured 30–40% visibility gains from statistics, citations, and quotations, and this is the section where they live. Section 5 is the differentiator most content skips: stating where your own advice fails ("if your site is client-rendered, this changes because several AI crawlers don't execute JavaScript") signals the first-hand expertise a language model cannot fake — which is exactly why models cite it.

How do you pick FAQ questions?

Apply one test to every candidate: would a person who landed on this page from this query actually type this question next? A post on ESOP taxation earns "when exactly does the tax event happen?" and never "what is a stock option?" — that second reader is on a different intent path and belongs on a different page. Off-intent FAQs don't just waste space; they blur what the page is about, diluting both its ranking focus and its answer-box eligibility. Three to six questions that pass the test beat ten that pad.

Edge cases and honest caveats

The framework fits question-shaped content — guides, definitions, comparisons, how-tos — which is most of what earns AI citations (informational content is cited at roughly 5x the rate of transactional pages in GeoSource's data). It fits poorly on news posts (lead with the event), pure tools and calculators (lead with the tool), and brand storytelling. And the framework cannot rescue a page nobody can retrieve: if crawlers are blocked or rankings are absent, structure is the second problem, not the first — the AI SEO operating guide covers that order of operations.