Verified against ChatGPT · 2026-08-08
Build a content brief from a target keyword and three competitor URLs that a freelance writer can actually execute without asking you ten questions
Produces a writer-ready brief — outline, angle, must-cover subtopics pulled from what's actually ranking, and word count rationale — instead of a generic bullet-point template that leaves the writer guessing what makes this piece different.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are building a content brief for a freelance writer who has never seen our site before and won't get to ask follow-up questions before drafting.
TARGET KEYWORD
how to switch payroll providers mid-year
TOP-RANKING COMPETITOR URLS
gusto.com/blog/switching-payroll-providers, adp.com/resources/mid-year-payroll-switch, rippling.com/blog/change-payroll-mid-year
OUR ANGLE OR DIFFERENTIATOR
We build the actual data-migration checklist ourselves for every customer, so we can go deep on the specific mid-year tax-filing gotchas that generic guides wave past.
AUDIENCE
HR managers at 50-200 person companies who've decided to switch but are nervous about the tax-filing transition specifically.
STEP 1 — READ THE COMPETITION
Treat the competitor URLs as representing what's currently satisfying this query, and infer their likely structure, subtopics, and depth from the keyword and URLs given (note explicitly that you're inferring rather than having crawled them live, and ask me to paste key excerpts if precision matters more than speed). List the subtopics that appear to be table-stakes — if all three competitors cover something, our piece has to cover it too or it reads as incomplete by comparison.
STEP 2 — FIND THE GAP
Identify one or two subtopics or angles that are plausibly under-covered given our stated differentiator, and explain specifically why our angle lets us go deeper there than a generic competitor would.
STEP 3 — BUILD THE BRIEF
Write the actual outline as H2/H3 headings, not a description of what an outline should contain. For each H2, give the writer one or two sentences on what needs to be said, not just the heading — a heading alone forces the writer to invent the content, which defeats the purpose of a brief. State a target word count per section, not just a single total, since a writer without section-level guidance almost always front-loads early sections and rushes the ending. Specify the primary keyword placement (title, one H2, first 100 words) and up to three secondary keywords to work in naturally, without instructing the writer to force them.
WHAT NOT TO DO
Do not write the brief as a set of generic content-writing best practices ("use short paragraphs, add a call to action") — every instruction in the brief must be specific to this exact piece. Do not invent statistics, studies, or named sources to include — if the brief calls for a supporting stat, instruct the writer to find and cite a real current source rather than presenting a placeholder as if it were factual.
OUTPUT FORMAT
1. Table-stakes subtopics (bullet list).
2. Content gap and why our angle earns it (one paragraph).
3. Full outline with word counts per section and one-to-two-sentence guidance per H2.
4. Keyword placement notes.
5. Suggested title and meta description (under 60 and 155 characters).Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
A content brief fails as a deliverable the moment a writer has to invent what a heading actually means, which is the default output shape GPT-5.1 reaches for when asked for "an outline" without being told to attach guidance sentences to every H2 — headings alone are cheap to generate and look complete, but they push all the real thinking back onto the writer, defeating the point of commissioning a brief at all. Explicitly separating table-stakes subtopics from the content gap forces the model to reason in two distinct passes instead of blending them, which matters because an LLM asked directly for "what should this piece cover" tends to merge "what's expected" and "what's differentiated" into one flat list, and a writer reading that list can't tell which sections are competitive necessities versus the actual reason to publish the piece at all. Instructing the model to flag that it's inferring competitor structure from URLs rather than having crawled them live matters because ChatGPT cannot actually browse the linked pages in this workflow unless browsing is explicitly enabled and used, and a brief that presents inferred structure as verified fact would send a writer chasing subtopics that may not actually be on the page — naming the inference honestly lets the requester decide whether to paste real excerpts before the writer starts. The explicit ban on inventing statistics addresses a known hallucination risk in brief-writing specifically: a plausible-sounding placeholder stat left in a brief has a real chance of surviving into the published piece if a writer skims past the caveat.
What you get back
H2: What actually breaks when you switch payroll mid-year (150 words) — cover the two tax-filing gaps generic guides skip: W-2 reconciliation across two providers and state unemployment insurance rate carryover. Primary keyword in title and first 100 words; secondary keywords "mid-year payroll transition," "payroll provider switch checklist" worked into H2s naturally.
Verified against
ChatGPT GPT-5.1 · 2026-08-08
Changelog
- 2026-08-08 — Initial publish, verified against ChatGPT GPT-5.1.
Building this for real?
This is a free starting point. If you'd rather have SEO built and running for your business, that's Scult's day job.
EXPLORE SEO
