The brief is the article: six fields that decide the quality of a draft

Six fields, in order

Topic is what the article is about, in one sentence a search engine could act on. Angle is what it says about it: the question it answers, or the reader it serves. Audience decides vocabulary and how much to explain. Language decides which outlets are searched first. Tone is a constraint on the writer, checked by a second model. Length sets the credit cost and the depth of the source set.

Every one of those changes something in the pipeline. None of them is decorative. A brief that leaves angle blank produces a survey; a brief that names it produces an argument, and arguments are what get cited.

Briefs that produce drafts nobody rewrites

The briefs that come back clean share a pattern. They name a question a real reader asks, in the reader's words. They say who the reader is. They give one or two URLs the writer must use, and they say what the article should end with: a checklist, a comparison, a decision.

The briefs that come back needing a rewrite are the ones written as keyword strings. "AI content SEO 2026 guide" is a search query, not a brief. The pipeline will find sources for it, and the draft will read like the sources: general, and about nothing in particular.

What the form enforces

The brief form in Brenzuri is those six fields and nothing else, plus optional URLs to include or domains to exclude and a brand profile for house style. It takes about two minutes. The estimate it shows before a run, in credits and in sources found, is the first feedback on whether the brief is good: a topic that finds two origins will produce a draft full of underlines, and the form says so before anything is spent.

Frequently asked questions