ChatGPT Content Brief Prompt Template for SEO Writers

A content brief is only useful if the writer can execute from it without back-and-forth. That means the brief needs a target keyword, a clear angle, a defined audience, a suggested structure, and at least a few notes on what competing content is already doing. Getting ChatGPT to produce all of that — in one pass, in a format your team can actually use — requires a prompt with more specificity than most templates provide.

This entry covers a tested prompt template for generating SEO content briefs using ChatGPT, with notes on which variables to fill in, what the model tends to miss, and where you need to apply editorial judgment before handing the brief to a writer.

What this prompt is designed to produce

The goal is a brief that a writer can open and start from — not a list of generic section headers. A well-formed ChatGPT output for this task should include:

• Primary keyword and 3–5 supporting semantic terms
• Search intent classification (informational, commercial, navigational, transactional)
• Target reader description — role, pain point, what they already know
• Recommended article angle and what to avoid (based on what's already ranking)
• Suggested H2 structure with 1–2 sentence notes per section
• Word count range with reasoning
• Content type (guide, comparison, tutorial, listicle, etc.)
• Internal/external linking suggestions if you supply context
• A short note on what the writer should verify or research independently

ChatGPT can produce all of this reliably when the prompt is structured correctly. The common failure is under-specifying the context — particularly the audience and the competitive angle. Without those, the model defaults to generic outlines that could describe any article on the topic.

The prompt template

Copy the block below. Replace everything in brackets with your actual values before submitting. Do not leave placeholder text in the prompt — the model will treat it literally and produce correspondingly vague output.

You are an experienced SEO content strategist. Your task is to produce a structured content brief for a writer.

Target keyword: [PRIMARY KEYWORD]
Business/site context: [1–2 sentences describing the site, its audience, and its positioning]
Target reader: [Who is this person? What role, what problem are they trying to solve, what do they already know?]
Content type: [guide / comparison / tutorial / listicle / landing page — pick one]
Competitor angle to avoid: [Describe what the top-ranking articles are doing that you want to differentiate from]
Tone: [Professional / conversational / technical / neutral]
Approximate target length: [e.g., 1,200–1,600 words]

Produce a brief that includes:
1. Primary keyword and 4–6 semantic/supporting terms to use naturally
2. Search intent classification with a one-sentence explanation
3. Recommended article angle — what specific perspective or framing makes this useful
4. Suggested H2 outline with 1–2 sentence notes per section explaining what to cover and why
5. Word count recommendation with brief reasoning
6. What the writer should NOT include or should actively avoid
7. 2–3 things the writer should verify, research, or source independently before writing
8. Optional: suggested internal linking opportunities if I describe related content

Format the output as a structured brief, not as prose. Use clear section labels.

Field-by-field notes

Business/site context

This is the most frequently skipped field. Without it, the model has no way to calibrate tone, depth, or angle for your specific audience. A sentence like "B2B SaaS company targeting operations managers at mid-market logistics firms" changes the brief substantially compared to "marketing blog." Two sentences is enough — you're not writing a brand guide, you're giving the model a positioning anchor.

Competitor angle to avoid

This field does real work. If you tell the model that the top results are all "beginner explainer" articles, it will steer toward something more differentiated. You don't need to name specific URLs — describing the pattern is sufficient. "Most ranking content is a generic step-by-step with no real examples" is enough context for the model to suggest a different angle.

If you haven't done any SERP research, leave this field out rather than guessing. The model will produce a reasonable default angle, which you can then revise after you've actually looked at what's ranking.

"Things to verify" section

This is the most practically valuable part of the output. ChatGPT will flag areas where it knows its output is likely to be shallow or outdated — pricing, statistics, platform-specific details, regulatory information. Treat this list as a mandatory pre-writing checklist, not a suggestion. If the writer skips it, you will almost certainly publish something with at least one factual problem.

What the model does well vs. where it falls short

An example filled-in prompt and what it produces

To make this concrete: here's the same template filled in for a real use case, followed by notes on what the output looked like.

You are an experienced SEO content strategist. Your task is to produce a structured content brief for a writer.

Target keyword: email deliverability checklist
Business/site context: Email marketing platform targeting e-commerce brands doing $1M–$20M in annual revenue. Audience is hands-on email marketers, not IT teams.
Target reader: An email marketer at a mid-size DTC brand who has seen deliverability drop after a list growth push and wants a practical fix list, not theory.
Content type: checklist/guide hybrid
Competitor angle to avoid: Most ranking content is technical (SPF/DKIM/DMARC setup) and assumes the reader is managing their own mail server. Our audience uses ESPs like Klaviyo or Mailchimp and needs platform-level guidance.
Tone: Direct, practical, assumes some email marketing experience
Approximate target length: 1,400–1,800 words

Produce a brief that includes:
[...same instructions as template above...]

With this level of specificity, GPT-4o produced a brief with a clearly differentiated angle ("deliverability for ESP users, not server admins"), a sensible 7-section H2 outline, and a "verify" list that correctly flagged Klaviyo's sending reputation tools and current Gmail/Yahoo bulk sender requirements as things the writer should confirm before publishing. The semantic terms it surfaced — sender reputation, list hygiene, engagement rate, bounce rate thresholds — were all on-target.

Without the competitor angle field, an earlier test run produced an outline that led with SPF/DKIM setup — exactly what we said we didn't want. The field matters.

Known failure modes

Adapting the template for different content types

The base template works for most informational and commercial-investigation queries. For specific content types, adjust the "Content type" field and add one instruction to the output list:

Using the brief output in your workflow

The brief ChatGPT produces is a starting point, not a final document. Before sending it to a writer, a content lead or SEO should do three things:

1. Run the primary keyword through your SERP tool and confirm the intent classification matches what's actually ranking. If the model called it "informational" but the top results are comparison pages, revise the angle accordingly.
2. Check the "verify" list and either resolve those items yourself or flag them explicitly in the brief so the writer knows what to source. Don't pass an unresolved verify list to a writer and expect them to handle it — they often won't.
3. Trim the outline if it's over-structured. An 11-section H2 outline for a 1,400-word article will produce a padded, shallow piece. Cut sections that overlap or that add length without adding value.

Prompt versioning note

This template was tested against GPT-4o in May 2026. OpenAI updates model behavior on an ongoing basis, and prompt outputs can shift meaningfully across model versions — particularly in how the model handles the "competitor angle" and "things to verify" sections. If you're running this on a different model (Claude 3.5 Sonnet, Gemini 1.5 Pro, or a future GPT-4o checkpoint), run a test brief on a topic you know well before deploying it across your content pipeline.

The structural logic of the prompt — give the model a specific audience, a competitive context, and an explicit output format — holds across models. The exact field labels and instruction phrasing may need minor adjustment depending on how a given model responds to imperative versus descriptive instructions.