Why this exists
The same audit of my client folder that exposed the missing briefs exposed something worse: for the projects that were underway, the discovery answers existed, but they were scattered across email threads, meeting notes, and my own memory. Who the client wants to sell to was in one email. The competitor they hate was said on a call and written nowhere. The budget lived in a chat message. Every one of those is a scoping decision, and none of them was in a document.
A project priced before discovery is a project scoped by hope, and scattered discovery is just slower hope. So this skill runs the interview properly and writes two documents at the end: the brief, which becomes the project's source of truth, and a strategy starter, which is the first thinking pass (a draft positioning sentence and two or three direction hypotheses). It pairs with my client-intake-builder skill, but they point in opposite directions: that one generates a client-facing HTML form the client fills in alone; this one interviews you (or processes the client's answers) and writes the documents.
How it works
- It reads before it asks Existing briefs, intake form answers, emails, meeting notes, the client's website. Anything the material already answers is never asked again, which is the difference between an interview and an interrogation.
- It interviews across seven areas, in small batches Business, audience, competitors, personality, style direction, deliverables, and timeline and budget. A few questions at a time, not twenty at once, and budget can be mapped against your own tiers if you keep a pricing config.
- It writes the brief and the strategy starter The brief holds every answer organised under the seven areas, keeping the client's own words where they are vivid. The strategy starter drafts a positioning sentence (for whom, the brand is what, unlike whom, because why), an audience snapshot, the competitive gap, and two or three direction hypotheses.
- It reports the thinking, not just the files The positioning sentence and direction hypotheses land in chat with both file paths, plus the natural next step: usually a proposal or moodboards. Unanswered questions stay listed as open items, never filled with guesses.
Step by step (for first-time users)
-
Type
/discovery-briefand name the project Something like "run discovery for the roastery rebrand". One project per run. - Point it at whatever already exists Intake answers, an email thread, call notes, the client's site. The more it reads, the fewer questions you get.
- Answer the interview in batches Expect short rounds across the seven areas. Answer in the client's words where you can remember them; vivid quotes survive into the brief.
- Read the strategy starter first The positioning sentence and the direction hypotheses are where you will disagree, and disagreeing now is the point: it is the cheapest moment in the whole project to be wrong.
- Work the open questions list Everything discovery did not answer is written down, not papered over. Take that list into the next client call.
The seven discovery areas
| Area | What it pins down |
|---|---|
| Business | What the client does, how they make money, what stage they are at |
| Audience | Who buys now, who they want to buy, price positioning |
| Competitors | Two or three names, and what the client envies or hates about each |
| Personality | Adjectives the brand should own, and the ones it must never be |
| Style direction | References they love, styles and colours that are off-limits |
| Deliverables | Exactly what the project produces, item by item |
| Timeline and budget | Deadline and range, mapped against your own tiers where possible |
Honest take
What it does well: The structure and the honesty. Seven areas asked in order means the awkward questions (budget, the competitor they hate) actually get asked instead of skipped, and reading the existing material first means the client never repeats themselves. The strategy starter is the part I did not expect to value this much: getting a draft positioning sentence and direction hypotheses minutes after discovery keeps momentum, even when the first draft is wrong, because a wrong sentence you can argue with beats a blank page you cannot.
What it does not do: It does not talk to your client. It interviews you, or processes answers the client already gave; for a form the client fills in alone, use client-intake-builder instead. It will not fabricate an answer to make the brief look complete: unknowns go under open questions and stay there until a human answers them. And the strategy starter is a starter, not the strategy; it is the first pass, meant to be interrogated.
When to use it: Before pricing, every time. Run it the moment a project is likely, so the proposal is built on answers instead of hope. It slots naturally between two of my other skills: client-workspace-scaffolder creates the folder with an empty brief, this fills the brief, and proposal-writer prices what discovery scoped.