What it means
When an AI tool reads a page to answer a question, it usually doesn't use the whole page. It splits the page into chunks, a paragraph or a section each, and picks the few chunks that best match the question.
Only those chunks go into the AI's answer. The rest of the page might as well not be there.
Why it matters
A long page can rank well and still give the AI nothing useful, if the answer is spread across many paragraphs or buried in a story.
And a single clear section can get quoted even if the rest of the page is about something else.
How to write for it
- One idea per section, with a heading that says what it is.
- Answer first, then explain. Put the key fact in the first sentence under each heading.
- Make each section stand on its own. Avoid "as mentioned above" or "it" referring to something three paragraphs back.
- Use tables and lists for prices, features and comparisons. They chunk cleanly.
- Name things. Write "Acme's Pro plan costs $29 a month", not "this plan costs $29".
Example
A pricing page has one long paragraph that mentions each plan in passing. An AI asked "how much does Acme cost?" picks a chunk that mentions only the free plan, and says Acme is free.
The company rewrites the page with a heading per plan and the price in the first line of each. Now the chunk that gets picked has every price in it.
Frequently asked questions
How big is a chunk?
It depends on the tool. Usually a few hundred words or less, often split at headings.
Does this mean long pages are bad?
No. Long pages are fine if each section is clear and self-contained. A long page just gives the AI more chunks to choose from.
Is chunking the same as RAG?
It's part of it. RAG is the whole process of fetching information before answering. Chunking is how the fetched pages get split up.