The First 200 Words: Why AI Engines Only Read Your Opening (And How to Write It)

Quick answer: Answer-first content places a complete, self-contained answer in the opening passage of a page, because AI engines retrieve and cite passages rather than whole pages. Deliver the full answer inside roughly the first 200 words, before any context, background or preamble. The Princeton and Georgia Tech GEO study found that adding citations, quotations and statistics to that opening lifts visibility in generative engine responses by up to 40%.

The opening is now the whole page

Pew Research tracked 68,879 Google searches from 900 US adults and found that users clicked a traditional result in 8% of visits where an AI summary appeared, against 15% where none did. Only 1% clicked a link inside the summary itself.

Those numbers reframe the job. When a machine reads your page and a human reads the machine, your ranking matters less than whether your words survive the summary. Answer-first content is the discipline that decides this, and most pages fail it in the first paragraph.

This post covers why retrieval systems only ever see part of your page, where the 200 word budget actually comes from, how to write an opening that gets quoted verbatim, and the three failures we find in almost every audit.

Why AI engines never read your whole page

AI engines never load your full page into a model. Retrieval systems split documents into passages, embed each passage separately, and pull back only the few that match a query. Your page is not a unit of retrieval. Your paragraphs are.

Retrieval systems index passages, not pages

Modern answer engines run on retrieval augmented generation, where a search layer fetches candidate passages and a language model composes an answer from them. We covered the full pipeline in our guide to retrieval augmented generation.

The practical consequence is blunt. A brilliant explanation in section four competes as an independent fragment against every other fragment on the internet, stripped of the title, the headings above it, and everything that gave it meaning.

The opening passage carries the page's topical identity

The first passage is the only one that reliably inherits what the page is about. It sits adjacent to the title and the H1, it contains the subject in plain words, and it survives every truncation strategy. Later passages inherit nothing automatically.

Google's own documentation sets the gate. Search Central states that to appear in AI features, "a page must be indexed and eligible to be shown in Google Search with a snippet". Snippet eligibility is passage eligibility, and the opening is your highest probability passage.

What answer-first content actually means

Answer-first content is writing that resolves the reader's question completely in the opening passage, then uses the rest of the page to prove, qualify and extend that answer. The answer does not build. It lands, then earns.

Journalists will recognise a cousin here, but the inverted pyramid is not the same thing. The inverted pyramid front-loads the most newsworthy fact. Answer-first content front-loads a self-contained response to a specific question, written so it stays accurate when someone removes every other word on the page.

Where the 200 word budget comes from

No search engine publishes its chunk size, and anyone quoting an exact figure from Google is guessing. The 200 word budget is a working constraint, not a documented threshold, and it is worth understanding why it holds.

Retrieval pipelines commonly chunk documents into a few hundred tokens with a small overlap, which lands near 150 to 250 English words. Write your answer inside that window and it occupies one clean passage. Spill past it and the answer splits across a boundary, arriving as two half-answers that each score worse than one whole one.

Treat 200 words as a budget you spend, not a target you fill. If the answer takes 60 words, stop at 60.

How to write answer-first content in five steps

Answer-first writing is a sequence, and the order matters more than the prose. Across the AEO audits we run at Loop2Tech in Karachi, Pakistan, the most common failure is not thin content. It is a strong answer buried under three paragraphs of throat-clearing.

1. Write the question exactly as a person types or speaks it. Not the keyword, the question. "Why is my page not cited by ChatGPT" beats "chatgpt citation optimization".
2. Answer it in two or three sentences, with the subject named. Never open with "This means" or "They pull from". Open with the noun.
3. Attach one piece of evidence to the answer. A number, a named system, a documented claim. The GEO researchers tested exactly this and reported gains up to 40%.
4. Delete the introduction you were going to write. We stopped writing them entirely. Context belongs after the answer, where it costs nothing.
5. Reread the opening with the title covered. If it no longer makes sense, it will not survive retrieval either.

Step five is the one people skip, and it is the one that catches the most damage. Our guide to optimizing content for Google AI Overviews applies the same test at section level.

The failures that quietly kill retrieval

Three problems account for most of the citation gap we see, and none of them show up in a standard SEO audit. Rank trackers and crawl reports cannot detect any of them.

Orphaned pronouns break every passage after the first

Open a section with "They pull from structured data" and the passage means nothing on its own. Open it with "AI Overviews pull from structured data" and the passage stands alone. This single habit, applied to every section, is the highest leverage change in GEO writing, and it costs nothing.

Restating the subject feels repetitive when you read the page top to bottom. Retrieval systems never read your page top to bottom.

DOM order decides what lands in the opening chunk

Extraction reads the HTML source order, not the visual layout. A breadcrumb trail, a cookie notice, a share bar or a "related posts" block placed before the article in the DOM will occupy your opening passage even when CSS renders it elsewhere on screen.

<!-- Costs you the opening passage -->
<article>
  <nav class="breadcrumb">Home &gt; Blog &gt; AEO</nav>
  <aside class="share-bar">Share this post</aside>
  <p>Answer-first content places the answer first...</p>
</article>

<!-- Keeps it -->
<article>
  <p>Answer-first content places the answer first...</p>
  <aside class="share-bar">Share this post</aside>
</article>

Check this with Screaming Frog by rendering a page and reading the extracted text in source order. The gap between what you designed and what a parser sees is usually larger than expected.

Preamble consumes your highest value real estate

A 300 word history of search before the answer does not delay the answer. It replaces it. The opening passage gets embedded either way, and yours will be embedded as a page about the history of search.

How to test whether your opening earns citations

Test the opening directly rather than inferring it from rankings, because rankings and citations diverge. A page can hold position three and never appear in a single AI answer.

Run three checks. First, query the engines with the exact question and record whether your domain appears, which we break down in our guide to measuring GEO and tracking AI citations. Second, compare impressions against clicks in Google Search Console for query clusters where AI Overviews appear, since a rising impression count with flat clicks signals summarisation without attribution. Third, paste your opening passage alone into a chat interface and ask what the page is about.

The third check is the fastest diagnostic we have found. If the model cannot name your topic from the opening alone, no retrieval system will either. Getting this right is the foundation of getting cited by ChatGPT, and it compounds with the entity work covered in our guide to building a knowledge graph presence.

Structure is the multiplier, not the whole system

AI content structure determines whether a passage can be retrieved. Authority determines whether it gets chosen once retrieved. Both matter, and teams that fix only one wonder why nothing moved.

The distinction between optimizing for answers, generative engines and classic rankings is worth getting straight, and we mapped it in AEO vs GEO vs SEO. For the wider system that direct answer writing sits inside, our generative engine optimization guide covers evidence density, entity consistency and citation tracking together. Teams that want this run as an ongoing programme can see how we structure it under our AEO, GEO and SEO services.

Start with your top ten pages this week

Pull your ten highest-impression pages from Google Search Console. Read the first 200 words of each with the title covered, and mark every page where the opening fails to answer the question it ranks for. Rewrite those openings before you touch anything else on the page.

Then apply the pronoun rule to every section heading you own. Name the subject, every time. These two changes cost no development work and no new content, and they change what a retrieval system can do with pages you already published. If you want the audit run properly, our AEO, GEO and SEO team does exactly this. Tell us which pages are underperforming and we will show you what the parser sees.

Frequently asked questions

What counts as the opening passage of a page?

The opening passage is the first block of body content a parser encounters in the HTML source, typically the first 150 to 250 words after the H1. It excludes navigation and headers only if those sit outside the main content region in the DOM. Answer-first content places the complete answer inside this passage, because retrieval systems embed it as the passage that represents the page.

How do I rewrite an existing blog post to be answer-first?

Find the sentence in the existing post that most directly answers the title's question, often buried in the middle, and move it to the top as the first paragraph. Name the subject explicitly rather than using a pronoun, attach one piece of evidence beside it, then delete the original introduction. At Loop2Tech we rewrite openings before touching anything else, because it is the cheapest change with the largest effect on citation.

Is answer-first content the same as writing a featured snippet?

Answer-first content and featured snippet optimization overlap but are not identical. Featured snippet writing targets a specific SERP feature with a 40 to 60 word paragraph, list or table. Answer-first content targets retrieval itself, so the answer must stay accurate and self-contained when extracted by any system, including ChatGPT, Perplexity, Gemini and Google AI Mode, none of which use Google's snippet format.

Why is my page ranking on Google but never cited by ChatGPT?

Ranking and citation use different selection criteria. Google ranks pages using links, relevance and quality signals, while generative engines retrieve individual passages and favour ones that are self-contained, evidence-dense and unambiguous about their subject. A page can rank on strong domain authority while every one of its passages opens with a pronoun and cites nothing, which makes it unusable as a quotable source.

Should every page on my site start with a direct answer?

Every page that targets a question should start with a direct answer, which covers blog posts, guides, documentation and FAQ pages. Homepages, category listings and portfolio pages serve navigation rather than questions, so a direct answer opening adds little there. The test is simple: if a person could type the page's purpose as a question, answer-first content applies.