Editorial Brief Generator
A prototype for turning raw interviews and notes into structured editorial briefs
Explores whether interviews, transcripts and raw notes can be turned into structured editorial briefs without losing judgement, nuance or narrative quality. The aim is not to automate writing — it is to make the thinking before writing more consistent, reusable and scalable.
Prototype. The Editorial Brief Generator is in early testing. The workflow is functional; the prompt logic is being refined through real briefs.
The problem
Editorial quality is difficult to systematise without losing what makes it valuable.
The gap between a raw interview transcript and a finished editorial brief involves significant judgement: what is the central story, which details prove the argument, where are the gaps, what does the reader need to know first.
That judgement currently lives in the heads of individual editors and writers. It is inconsistent, unscalable, and lost when people move.
The question this experiment is testing: can editorial judgement be encoded into reusable workflows without reducing it to a checklist?
The experiment
The Editorial Brief Generator is a structured AI workflow that takes raw inputs — interview transcripts, customer notes, campaign context, narrative themes — and produces editorial briefs that are ready to brief a writer or develop directly.
The aim is not to write the content. It is to do the thinking before writing: identify the story, extract the evidence, structure the argument, surface what’s missing.
Inputs
- Interview transcript or customer notes
- Campaign context or narrative theme
- Relevant evidence and research
- Intended format (article, case study, sales story)
Process
Identify the central story
↓
Extract useful proof points
↓
Structure the brief
↓
Surface risks and gaps
↓
Recommend article angles
Each step is prompted explicitly. The workflow asks the model to make editorial choices — not just summarise — and to flag where the evidence doesn’t support the argument or where a follow-up question would strengthen the brief.
Outputs
- Article brief — headline, angle, key argument, evidence, structure
- Case study brief — customer situation, intervention, outcome, reusable insight
- Sales story notes — problem, solution, result, objection handling
- Follow-up questions — what the brief needs that the raw material doesn’t provide
- Reusable evidence — proof points that can travel across formats
What’s being tested
- Whether the same input produces consistent brief quality across different types of content
- How much the prompt design needs to vary by content type versus narrative theme
- Whether the briefs produced are actually useful to writers — shorter briefing time, clearer angle, fewer revision cycles
- Where human editorial judgement is still essential and cannot be encoded
What this is not
The Editorial Brief Generator is not a writing assistant. It does not produce finished content.
It is a thinking tool — a way of making the editorial reasoning that precedes writing more explicit, more consistent, and more transferable.
The experiment will have succeeded if it reduces the gap between raw material and usable brief without reducing the quality of the editorial judgement applied to that gap. Whether that is achievable at scale is still an open question.