Validate a startup idea the way a good founder would — then ship it.

Act one — Understand

The first four questions turn a raw idea into something specific enough to be testable. This is where most AI tools skip straight to generating code; Motriz starts where a founder would.

Question 1
What are we working with?

Get the lay of the land so you don't repeat work you've already done. Existing research, an existing repo, or nothing at all — the journey starts from your real situation.

Question 2
What's the idea?

Pin the idea down in your own words before it gets pulled apart. The co-founder runs a short intake: the idea in one sentence, who it's for, and what those people do about the problem today.

Question 3
Who has this problem?

Name the people who feel this pain so you know who you're building for. Problem-framing and customer-profile specialists sharpen the picture into a concrete ICP, not a demographic guess.

Question 4
What's already out there?

See who solves this today and where the real gaps are. A competitor map keeps you honest about why anyone would switch.

Act two — Validate

Assumptions get the cheapest tests that could kill them. Evidence — interviews, signups, experiments — hangs on the journey itself, not in chat scrollback.

Question 5
How will we test it?

Decide what would prove the idea right or wrong before spending real effort. Motriz designs a validation roadmap around the riskiest assumptions first.

Question 6
What did we learn?

Run the tests and face what the evidence actually says. Every result lands on the journey as evidence you can argue with — for the idea or against it.

Act three — Decide & Build

Question 7
How would this win?

Turn the evidence into a sharp plan for how this beats the alternatives — a strategy memo, not a vibe.

Question 8
What do we build first?

Scope the smallest first version that proves the next big assumption. The MVP scope is deliberately narrow: it exists to keep learning, not to demo well.

Question 9
Build it — or not?

The gate. The co-founder prepares the call — evidence for, evidence against, open risks — and returns one of four honest verdicts: Build, Validate more, Pivot, or Kill. You decide, and every verdict is recorded with its reasoning. If the answer is stop, you just had the cheapest failure of your career.

After the gate: the handoff

When the verdict is build, the venture becomes a code project. Four artifacts travel into your repository — the PRD, the MVP scope, the UX flow, and an evidence pack digesting the interviews, experiments, and gate decisions that justified the build. Raw documents never enter the repo; the co-founder keeps the full memory.

The MVP scope seeds the first tasks, queued riskiest-first. Coding agents — Claude Code, Codex, or Ollama Cloud, running on the subscriptions and keys you already have — execute them one at a time or in parallel isolated Git worktrees, with plan → execute → review flows that loop until the review approves. Nothing runs until you press Run, and every task opens already knowing the venture: the evidence, the decision, the spec.