/visionaire:startThis is the hour that prevents the month of building the wrong thing.
Convert vague intent into a precise, validated definition. Specific problem. Real user. Measurable success. Explicit scope — before any building begins.
None of this is theoretical. These are the patterns that show up every time.
The Product Brief is where you find out you were solving the wrong problem — before you've written a line of code. Skip it, and every downstream phase executes with conviction on a definition nobody wrote down and everyone interpreted differently.
CB Insights analyzed 101 startup failures. 42% cited "no market need" — not bad execution, but building the wrong thing because the problem was never precisely defined.
The Product Brief Phase takes what lives in your head — specific, contextual, complete — and extracts it into a written definition everyone can execute on.
The adaptive interview doesn't follow a script. It follows what you reveal — and it keeps asking until the gaps are closed.
Market Validation validates what the Product Brief defined. UI/UX designs for who the Product Brief specified. Architecture builds what the Product Brief scoped. Ambiguity here multiplies through everything that follows.
Start my spec session — $297The outline-first pattern separates understanding from writing. The structure is reviewed and approved before any prose is written.
The agent reads IDEA.md and presents its understanding of what was explored — what excited you, what you assumed, what you decided not to build. You confirm or correct. The interview starts from shared understanding, not the agent's interpretation.
One question at a time, drawing from six categories: Problem & Motivation, Target User, Desired Outcome, Market & Alternatives, Differentiation, Success Criteria. Vague answers get follow-ups. Concrete answers unlock the next layer.
Before writing anything, the agent reviews IDEA.md again — surfacing assumptions not yet addressed, open questions unanswered, variations unaccounted for. Gaps are resolved here, not discovered in the Brief Review stage gate.
BRIEF-outline.md is generated showing what was understood, what assumptions the agent is making, and what's missing. You approve, request corrections, or keep refining. This loop repeats until the structure is right — before a word of prose is written.
Once the outline is approved, BRIEF.md expands it into 2–4 pages — problem, users, use cases, goals, success criteria, constraints, scope in and out. Then Brief Review evaluates it against eight criteria before any design work begins.
Unlike the Idea Phase, the Product Brief has a formal stage gate. An autonomous review agent evaluates BRIEF.md against eight criteria — problem clarity, target user specificity, desired outcome, internal consistency, business model presence, evidence basis, founder context, and stability for design. APPROVED only if all eight pass. REJECTED triggers revision before design begins.
The adaptive interview in practice
When your answers are vague, the interview keeps asking. This is what that looks like:
Three ways the Product Brief Phase finds its footing — including when the interview is over and the document comes back wrong.
The interview felt productive. Then the outline arrives and it's not quite right. The problem statement focuses on time savings when you meant something about trust. The use case misses the key context. You'd need to rewrite half of a full document.
The outline-first pattern catches this at minimal cost. You review the structure — not prose, just what was understood — and correct it before any writing happens. Two minutes, not twenty.
A revised outline that correctly captures what you meant — before any prose investment. The document that emerges reflects your actual intent, not the agent's first interpretation.
You approved the outline. You approved the full document. Brief Review comes back: REJECTED. "Problem statement conflates two different problems. Success criteria undefined. Scope unclear between v1 and future." You're frustrated — you thought you were done.
The rejection isn't failure, it's precision. The review found what would have caused problems in design: a designer would have guessed which problem to solve. The rejection includes exactly what to fix and why. Answer three specific questions, the agent revises the document, the second review returns APPROVED.
BRIEF-v2.md that passes all eight criteria — a document that provides design stability, not one that looks complete but leaves fundamental questions open.
Brief Review rejects with a finding you didn't expect: "The brief conflates three different problems — search, trust, and learning. These may be separate products." You look at the rejection and realize the reviewer is right. You've been conflating three ideas.
The orchestrator offers to return to the Idea Phase with the Brief Review feedback as input. The Idea Phase re-opens, this time focused on which of the three problems to pursue. A refined IDEA-v2.md emerges. The Product Brief Phase restarts from the clarified foundation. The iteration loop worked exactly as designed.
A Product Brief built on a genuinely clarified idea — not one that tried to specify its way through an ambiguous foundation. The review loop caught what would have caused problems for every phase that followed.
Creative exploration ends here. Disciplined definition begins. Every phase that follows executes on what the Product Brief establishes — not their own interpretation of what you said in a kickoff meeting.
The Product Brief is the first phase with a formal stage gate. If Brief Review rejects, the workflow halts. Nothing advances on an ambiguous foundation.
Teams do it. The tradeoffs are the same every time.
Product Brief clarity compounds. Market Validation validates what the Brief defined. UI/UX designs for who the Brief specified. Architecture builds what the Brief scoped. Features implement what the Brief committed to. Ambiguity in the Product Brief multiplies through every phase that follows. Clarity in it does too.
I built the outline-first pattern after watching teams approve interviews that felt right but produced documents that missed the point. The adaptive interview doesn't follow a script — it follows what you reveal. And the stage gate isn't courtesy. It enforces what your intuition alone can't.
— Robert EvansProduct Brief misalignment is cheap to fix before design starts. After architecture and features build on it, it costs weeks — not a conversation.