For about three years my content process started the same way every single time: I would open six competitor tabs, copy their H2s into a spreadsheet, stare at the gaps, then write a brief that smooshed the best of everything together with my own ideas on top. When I was a solo freelancer billing by the hour, this was honestly fine. Then I joined an agency that ships 15 to 20 articles a week, and my beautiful artisanal half-day brief became a bottleneck with a heartbeat. The writers were waiting on me. I was the slow part. Nobody wants to be the slow part.
The thing that actually fixed it was not 'do the same process but faster.' It was the Search-Intent Content Brief — Turn One Keyword Into a Writer-Ready Outline prompt, and specifically the way it refuses to start with structure. It starts with intent. That reordering sounds like a small thing. It is not a small thing. It is the whole game.
The day I realized I'd been writing how-to guides for shopping queries
Here is an embarrassing confession. For a keyword like 'best CRM for small business,' I had, more than once, briefed a 'how to use a CRM' article. The intent there is Commercial Investigation — the person is comparing options with a credit card half out of their wallet — and I was handing them a tutorial. We ranked at position 11 and I could not figure out why a 2,400-word guide wasn't moving. It wasn't moving because it answered a question nobody was asking. When I ran that exact keyword through this prompt, Claude Opus 4.8 labeled it Commercial Investigation in the first line, recommended a comparison format, and quietly made me feel like I'd been doing this job with my eyes closed.
That single correct call is worth the entire brief. Everything downstream — the outline, the snippet flags, the FAQ — is built on the intent hypothesis, so getting intent right is the load-bearing wall. Claude is genuinely good at this. The XML-ish tags in the body (<role>, <task>, <output_format>) keep it disciplined, and the 'think step by step, then output only the finished brief' instruction means I get clean output instead of a model narrating its feelings about CRM software for four paragraphs first.
The parts of the output I actually use
The [SNIPPET] flags are the most immediately useful thing in the whole brief. Each flagged section, I tell the writer to nail a definition or a tight numbered list inside 100 words, because that is the shape Google lifts into a featured snippet box. The FAQ block comes pre-phrased the way real humans type questions, which means it doubles as my FAQ-schema source later. And the internal-link placeholders are sneakily valuable: on our last batch, three of the five suggested anchors pointed at pages we did not even have yet. That is not a brief. That is a content gap map wearing a brief costume.
- Intent classification — the one section every writer reads before touching the outline
- [SNIPPET] flags — tells writers exactly where to slow down and be surgical
- FAQ block — phrased as typed queries, reused later for FAQ schema
- Internal-link placeholders — quietly surface pages you haven't built yet
- Word-count + pillar-vs-supporting call — kills the 'is this a 900 or a 2,500 word piece' debate
The honest caveat
The brief is only as good as the intent call, and the model is not always right. About once a month it picks the wrong intent on a genuinely ambiguous keyword — usually something that sits between informational and commercial. So I added a 30-second habit: before any writing starts, I glance at the live SERP and sanity-check the intent label against what is actually ranking. That tiny check has caught two misclassifications in the last few weeks. The prompt even hedges its own guesses with [VERIFY ON SERP] when no competitor URLs are supplied, which I appreciate — a tool that tells you where it is unsure is a tool you can trust with the parts where it is confident.
These days my workflow is: run the brief, do the 30-second SERP check, then I pair it with my Keyword Cluster Builder — Group Seed Keywords Into Topical Themes prompt to make sure the article slots into an actual cluster instead of floating alone. The two together turned content planning from a guessing ritual into a repeatable assembly line. Grab the Search-Intent Content Brief — Turn One Keyword Into a Writer-Ready Outline prompt on Prompt Dock and run it on your own next keyword — the moment it corrects an intent you had wrong, you'll get it.