Writing answer-first content that gets quoted
where AI assistants pull quotes from on a page, and how to write the 40 to 60 word answers that get cited instead of skimmed past.
last updated 30 October 2026
if you want an assistant to quote you, put the answer at the top. Kevin Indig’s study of 1.2M AI answers found that 44.2% of ChatGPT citations come from roughly the first 30% of a page. so lead with the answer, put a 40 to 60 word standalone answer under every heading, and let the explanation follow instead of preceding.
Why the top of the page wins
assistants extract passages, not pages. when a model retrieves your page it is hunting for a span of text that answers the user’s question on its own, and the evidence says it finds that span early or not at all. 44.2% of citations from the first ~30% of the page, in a sample of 1.2M answers, is not a subtle signal. every paragraph of throat-clearing you put above the answer, the backstory, the “in today’s landscape” opener, the definition of terms nobody asked about, pushes your quotable text toward the part of the page that rarely gets quoted.
this is the opposite of how most content marketing was taught. the old model held the answer back to keep you scrolling. the extraction model rewards giving it away in the first breath.
The 40 to 60 word answer block
under every heading, write one paragraph that answers that heading’s question completely, in 40 to 60 words, with no dependence on the text around it. that is the unit assistants lift. the test is simple: read the paragraph alone, out of context. if it holds up as a full answer someone could paste into a chat, it works. if it needs the previous section to make sense, it will not survive extraction.
the heading itself is half the block. phrase it as the question a buyer would actually type, not as a clever label, because retrieval matches the question before it ever reads your answer. a heading like “pricing” earns nothing. “what does x actually cost” is a hook the question can catch on.
this page is doing it in front of you. so is the rest of this site, which is built as an open GEO experiment, documented in how i run this site.
Fact plus interpretation beats narrative
the style that gets cited is a specific fact followed by what it means. “44.2% of citations come from the first 30% of the page, so put the answer first” is quotable. a winding story that arrives at the same conclusion in paragraph nine is not, because no single span of it stands alone. narrative writing spreads its meaning across the arc, and extraction kills arcs.
keep the numbers exact, name the source in the sentence, and attach one plain sentence of interpretation. that pairing is what an assistant can safely reuse: a claim it can attribute plus a takeaway it can hand the user.
Briefing prose beats guide prose
write like you are briefing a busy operator, not narrating an “ultimate guide”. short declaratives. one point per paragraph. headings that are real questions, phrased the way a buyer would type them. the sprawling 6,000 word guide voice was built to maximize time on page, and assistants have no time on page. clarity per span is the metric now, and briefing prose wins it without even trying to.
how assistants choose which products to name at all is a separate question, covered in how AI assistants actually pick products.
FAQ
Does answer-first writing bore human readers?
no, it respects them. humans skim exactly the way extractors do, and the reader who gets the answer in the first paragraph decides you are worth trusting with the details below it. the people bounced by a direct answer were never going to convert.
Should every page target one question?
yes, one primary question per page, phrased in the title the way a buyer would ask it. sections then take the follow-up questions. a page aimed at four questions at once gives the extractor four half-answers and gets cited for none of them.
How long should the page be?
length matters less than where the answer sits. 700 words with the answer in the first 100 beats 3,000 words with the answer in the middle. add length only when each section still passes the standalone test.
the rest of the method, both levers end to end, is in the AI search playbook. the breakdown call is free. Let’s talk Growth.
this page is part of the AI search & GEO playbook.
More in this playbook
How AI assistants actually pick products
ChatGPT recommends a tiny fraction of the products it knows. what separates the named from the ignored, and why it differs per assistant.
AI search & GEOReddit's role in AI answers, honestly
Reddit shows up in roughly 40% of AI citations and the share can drop overnight. how to use the channel without building on sand.
AI search & GEOThe technical layer: what matters and what is theater
static HTML, fast pages and visible answers move AI citations. schema and llms.txt mostly do not. what to build and what to skip.
AI search & GEOListicles and comparison pages AI loves to cite
why lists and comparison tables get cited so often, how to write an honest one, and how to get into other people's.
AI search & GEOMeasuring your AI share of voice
a free monthly prompt panel that shows which assistants name your product, which name competitors, and which sources they read.
AI search & GEOAI directories: six good ones beat sixty
a handful of real AI directories beat mass submission. what a listing actually does for AI visibility and how to submit cleanly.
AI search & GEOThe honest GEO timeline
expect 8 to 16 weeks before AI assistants start naming you. what compounds, what spikes, and why a young domain still has a shot.