How long should an intro summary be?

40–80 words. Short enough to be read as a standalone excerpt, long enough to convey topic, claim, and audience. Under 40 words is often too thin to establish full context; over 80 words risks becoming a second paragraph rather than a summary.

Should the intro summary contain the target keywords?

Yes, naturally. The intro summary should mention the primary topic keyword because it's describing what the page is about — not to satisfy a keyword density requirement. Write for the reader first: if the summary is clear and accurate, the keywords will appear naturally.

Does this advice apply to landing pages as well as articles?

Yes. Product landing pages should open with a clear value proposition: what the product does, who it's for, and what outcome it delivers. The same principle applies: front-load the most valuable information. AI assistants answering 'what does X product do?' read the first paragraph of the product page.

Get Started

Are your pages being cited by AI engines? Audit your GEO score for free.

Get a demo

Core GEO Signals · Published Mar 31, 2026

Intro Summary

A concise opening paragraph (under 80 words) that states what the page is about and who it's for.

TL;DR — AI engines read the first paragraph to decide page relevance before processing the rest of the content. A direct, 40–80 word summary at the top guarantees the conclusion is captured — even if the model never reads further.

Why Intro Summaries Matter for AI Engines

The first paragraph of a page carries disproportionate weight in how AI engines assess relevance. Before processing the full content, a model reads the opening paragraph to determine whether the page matches the user's query. A long, throat-clearing introduction — "In today's rapidly changing digital landscape, businesses face unprecedented challenges..." — wastes the model's attention budget without conveying any useful information.

The Princeton GEO paper (2024) analysed pages that appeared as cited sources across ChatGPT, Perplexity, and Google AI Overviews. Pages with clear, direct opening summaries — stating the topic, the key claim, and who benefits — were cited significantly more often than pages with indirect or vague introductions. The mechanism is direct: the model's relevance assessment runs on the first N tokens of a document. A direct summary front-loads the most valuable signal.

This also intersects with the "lost in the middle" problem in LLM research: information at the beginning and end of a document is weighted more heavily than information in the middle. Your first paragraph is the highest-attention-weight content on the page.

How to Implement

First paragraph: 40–80 words
State: what the page is about + the key claim + who benefits
Avoid passive voice and filler phrases
The opening paragraph should be self-contained — it should make sense as a standalone excerpt when extracted by an AI engine

Good example:

"Schema markup is JSON-LD structured data added to a page's <head> that tells AI engines exactly what type of content the page contains. Pages with valid Article or FAQPage schema are cited 30–40% more often in AI-generated answers (Princeton, 2024). This guide covers implementation for the five most important schema types."

Weak example:

"In this comprehensive guide, we'll be taking an in-depth look at schema markup and exploring how it can potentially benefit your website in various ways..."

Common Mistakes

Starting with a question — "Have you ever wondered why your content isn't being cited by AI?" delays the actual answer; start with the answer
Repeating the page title verbatim — the first sentence should add information beyond the title, not restate it
Generic openers — "Welcome to our guide on..." or "In today's digital landscape..." are irrelevant filler that erodes the model's attention budget