What is a content brief?

A content brief is the document a strategist hands a writer before a page is written. It defines the question the page answers, who it is for, what it must cover, which sources to use, and how it connects to the rest of the site.

How is an AI SEO content brief different from a traditional SEO brief?

A traditional brief is built around a keyword, a word count, and competitor headings. An AI SEO brief is built around the question and its likely follow-ups, a quotable one-sentence answer, sourced facts, and a passage plan, because AI engines select individual passages rather than whole pages.

Should a content brief include a target word count?

Only as a loose range. Length should follow from the number of questions the page needs to answer well. A word count target tends to produce padding, and padding makes passages harder to quote.

Who should write the content brief?

Whoever owns the strategy and knows the buyer, often with input from sales or support, who hear the real questions. The writer should be able to produce the page from the brief without guessing at intent.

Can I use AI to write the brief?

AI tools are useful for drafting fan-out questions and spotting gaps, but the facts, sources, and point of view in a brief should come from people who can stand behind them. A brief is where accuracy is decided.

The AI SEO Content Brief: A Template for Writers

An AI SEO content brief is a writer’s specification built around the question a page answers, the follow-up questions an AI engine will ask, and the facts that make the answer quotable. The traditional brief of keyword, word count, and competitor headings still produces pages that can rank. It rarely produces pages that get cited. Below is the template we use, how to fill in each field, and a worked example.

Why the old brief falls short

Traditional briefs optimize the page for a keyword. AI engines work differently in two ways that change what a brief has to specify.

They select passages, not pages. An engine retrieves candidate pages and then quotes the specific passage that best answers the question. A brief that does not plan passages leaves citability to chance. We explain the mechanics in content chunking for AI.

They answer questions nobody typed. Engines expand a prompt into many follow-up searches. An AirOps analysis of 15,000 ChatGPT prompts, reported by Search Engine Land in March 2026, found those prompts expanded into 43,233 queries, and 95% of the expanded queries had no traditional search volume. A keyword-led brief cannot see them. A question-led brief can.

What changed since the 2023 brief

When we checked the results for “SEO content brief template” in September 2026, all three top-ranking guides were published in November 2023, before AI Overviews rolled out widely in the US and before ChatGPT search launched. They are still good at what they set out to do. They just brief for a results page that has since changed underneath them. Field by field, this is what moved:

Brief field	2023 brief	2026 AI SEO brief
Target	One primary keyword plus secondaries	One primary question plus its fan-out questions
Opening	Hook, then context	The answer in one quotable sentence
Structure	Competitor H2s, rewritten	One section per question, each answering in its first line
Evidence	“Include stats”	Each figure with its source and date, listed in the brief
Length	Target word count from the top 10	Whatever the question set needs; no fixed count
Identity	Brand voice notes	The entity speaking, and which service the page supports
Upkeep	Rarely specified	A review date and a list of facts likely to go stale

The template

Copy this into your brief document. Every field has a job; if one is empty, the page will usually be weaker in exactly that spot.

AI SEO CONTENT BRIEF

1.  Working title:
2.  Primary question (as a buyer would ask it):
3.  The answer in one sentence (becomes the page's opening line):
4.  Audience and stage (who, what they know, what they decide next):
5.  Entity (who is speaking; which product or service this supports):
6.  Fan-out questions (5 to 10 follow-ups a buyer or an AI would ask):
    -
7.  Facts and sources (each: figure, source, date):
    -
8.  Passage plan (one H2 per question; first sentence answers it):
    H2:                    Answers:                    Evidence:
9.  Decision content (cost, time, who it suits, who it does not):
10. Comparisons or alternatives to address:
11. Internal links (up to the pillar, across to siblings, one service page):
12. Schema (page type; FAQ only if the Q&A is visible on the page):
13. Freshness (review date; facts likely to go stale):
14. Do not (claims we cannot support, tone, topics to avoid):

How to fill in each field

Primary question

Write it the way a buyer would say it to a colleague or type it into ChatGPT, not as a keyword. “How much does an SEO audit cost for a 50-page site” is a question; “SEO audit cost” is a keyword. The question defines intent; you can still check its search volume separately.

The answer in one sentence

This is the most important field on the brief. It becomes the first sentence of the page, and it should work as a standalone quote: specific, declarative, and true without the paragraphs around it. If the strategist cannot write it, the page is not ready to be written.

Fan-out questions

List what the buyer asks next. Three sources work well: ask ChatGPT, Perplexity, and Gemini the primary question and note the follow-ups they suggest; ask sales and support what prospects ask before buying; and read the “People also ask” box. Each fan-out question either earns its own section or its own page in the topic cluster.

Facts and sources

Every statistic gets a source and a date before writing starts, not after. Writers who have to find numbers themselves either skip them or grab the first one they find. Sourced, dated facts are also what make a passage safe for an engine to repeat.

Passage plan

One H2 per question, phrased as the question or as a clear claim. Beside each, write what the section must conclude and what evidence supports it. This is the outline, and it is also the citability plan.

Entity

State who is speaking and which service the page supports. It keeps the page anchored to your brand, which matters for entity recognition, and it decides the one commercial link the page carries.

Decision content

Cost, time, who it suits, and who it does not. Validation questions are the hardest to be cited for, and pages that answer them honestly, including “this is not for you if”, stand out because so few do.

Links, schema, and freshness

Links go up to the pillar, across to two or more siblings, and out to one service page. Schema is the page type plus FAQ markup only if those questions and answers are visible on the page, which is Google’s rule. Freshness is a review date and a note of which facts will age first.

Building the fan-out list in 15 minutes

Field 6 is the one writers most often leave thin, because it looks like research. It is quick if you work from four sources in this order and stop once new questions stop appearing.

People Also Ask and related searches for the primary question. Expand two or three PAA items; each one loads more.
The AI engines themselves. Ask ChatGPT, Perplexity and Gemini the primary question, then note what they cover that you did not plan for, and the follow-ups they suggest.
Sales and support. The questions buyers ask on calls, especially objections, are the fan-out questions nobody publishes answers to.
Your own cluster. Check which sibling pages already answer a question, so the brief links to them instead of repeating them.

Keep a question if a buyer would genuinely ask it before deciding. Drop it if it only exists because a tool listed it. Most pages end up with six to ten.

A worked example

Here is how the brief might look for a page answering whether a business should build an llms.txt file, the subject of our llms.txt guide.

1.  Working title: What Is llms.txt (and Is It Worth Building)?
2.  Primary question: Do I need an llms.txt file for my website?
3.  Answer: llms.txt is a proposed file that gives AI models a clean
    map of your best content; it is cheap to build, but the major
    engines have not confirmed they use it, so treat it as optional.
4.  Audience: marketers and developers who have heard the term and
    want to know whether to spend time on it.
5.  Entity: Ambeltek; supports the AI SEO service.
6.  Fan-out: Who proposed it? Does ChatGPT read it? How is it
    different from robots.txt and sitemap.xml? How do I build one?
    Can it hurt? What should go in it?
7.  Facts: proposed in September 2024 by Jeremy Howard of Answer.AI
    (llmstxt.org); engine support status as of publication date.
8.  Passages: What it is / Who proposed it / Do engines read it /
    llms.txt vs robots.txt vs sitemap / How to build one / Verdict.
9.  Decision: worth an hour for most sites; not worth delaying
    fundamentals for.
11. Links: up to the GEO guide; across to robots.txt and JS
    rendering posts; out to the AI SEO service.
12. Schema: BlogPosting; FAQ only for questions shown on the page.
13. Freshness: review quarterly; engine support may change.
14. Do not: claim any engine uses it without a source.

Notice that field 3 is already a quotable answer and field 14 prevents the most likely inaccuracy. Those two fields do most of the work.

Brief review checklist

Before a brief goes to a writer, check it against this list.

Where the content brief sits in the workflow

The brief sits between strategy and writing: the cluster plan decides which pages exist, and the brief decides what each page must do. If you are planning a cluster from scratch, start with the 90-day AI SEO roadmap. If you would rather hand the whole pipeline over, from question mapping to briefs to published pages, that is the core of our AI SEO service.