Analyst for Positioning Brief Generation

The blank Google Doc is the most expensive asset in product marketing. PMMs we work with lose two-to-three days per brief staring at it — not because they don't know the company, but because positioning prose has a tax: every sentence is reviewed by four people, and the cost of a wrong opening paragraph is a week of revisions.

An AI analyst flips that math. Not by writing the brief for you — that produces the slop you've already seen — but by collapsing the time between "I have inputs" and "I have a draft worth red-lining." The brief still ships from your judgment. The blank page is what disappears.

A note on terms before going further. By positioning brief we mean the internal document — usually four to eight pages — that codifies who you sell to, what category you compete in, and the proof points sales and marketing reference downstream. Not a launch announcement. Not the homepage. The source-of-truth artifact that messaging gets derived from.

The analyst's job is to remove the blank page, not to replace your judgment.

What the analyst is actually doing

A positioning brief drafted from scratch by an LLM is recognizable within ninety seconds — generic category nouns, hedged differentiators, customer quotes that sound suspiciously like marketing copy. The output fails because the input is wrong. You asked the model to invent positioning. It can't. It doesn't know your win rates, your sales objections, or which feature your three biggest customers actually use.

What it can do — well, and at speed — is structure inputs you already have. Sales call transcripts. Win/loss interviews. Competitor pricing pages. The half-finished Notion docs from last quarter's offsite. The analyst's value is synthesis under a positioning frame, not generation from nothing.

The inputs that change the output

Most of the difference between a useful draft and a generic one is upstream of the prompt. Before opening the analyst, gather:

If you can't assemble these in an afternoon, the brief isn't the bottleneck — the underlying research is. No prompt fixes that.

The four-prompt sequence

The mistake is asking for the whole brief in one prompt. You get a draft that sounds confident but disagrees with itself by page three, because the model committed to a category noun on page one before seeing the differentiation argument on page four.

Better to run the analyst through a sequence, where each step's output becomes the next step's constraint.

What you keep, what you rewrite

Even with clean inputs and a sequenced run, the draft has predictable failure modes. Plan to rewrite the same sections every time.

A worked example of the difference

The category noun in the final brief wasn't the analyst's. But the analyst surfaced the question that unlocked it — and that's the point.

What to do this week

If you have a brief on your roadmap for the next quarter, spend Monday assembling the input pack from the checklist above. Don't open the analyst yet. The leverage isn't in the prompt — it's in whether you have honest transcripts and a real read on competitors before you start.

By Tuesday afternoon, run the four-prompt sequence. Expect to spend Wednesday and Thursday editing the draft into something that sounds like your company. Friday is for review.

That's a one-week brief, where the previous baseline was probably four. The analyst didn't write it. It removed the blank page so you could.