← MVPAI in a Week Day 01 · 4 sessions

The route

DAY 01 · SESSION 01 · MORNING

Products start with problems

This week you take one idea along the whole MVPAI route. This morning starts where every product should: with a real person and a real difficulty.

You will be able to
  • Explain how a product differs from a project
  • Explain why AI makes building cheaper but does not make being wrong cheaper
  • Place any claim on four levels of evidence
  • Write a problem statement: user, situation, struggle, consequence, evidence
  • Describe the MVPAI route, how gates work and which checks must always pass

OUTPUT · A one-line problem statement, saved in product/01-route.md

DAY 01 · SESSION 02 · MORNING

From idea to MVP

Who it is for, what they use today, what you are betting on, and the smallest product that tests the bet.

You will be able to
  • Choose a narrow first segment and name today’s alternatives, including doing nothing
  • Write a value proposition and name the riskiest assumption behind it
  • Define the first value moment and the smallest MVP, with scope in and out
  • Describe a system in plain English: interface, logic, data
  • Explain what an ADR records and why

OUTPUT · A draft value proposition and first value moment in product/01-route.md

DAY 01 · SESSION 03 · AFTERNOON

Building with Claude

How to work with Claude as an engineering partner: briefed clearly, building in small loops, and checked by you before anything is kept.

You will be able to
  • Explain how Claude.ai and Claude Code differ, and when to use each
  • Brief Claude with context, goal, reasons, an example and what done looks like
  • Describe the build loop: plan, build, verify, review, commit
  • Ask for the right tests and read a diff with five questions
  • Explain server-side access control and where secrets belong

OUTPUT · A first brief for your slice, saved in product/01-route.md

DAY 01 · SESSION 04 · AFTERNOON

Launch, learn and decide

A demo is not a launch. This session covers your first real user, what it costs, and how to decide what happens next.

You will be able to
  • Tell a demo from a launch, and plan a first-user test
  • Estimate cost per user, including AI, and say which numbers are assumed
  • Write break-even in one line
  • Choose between persevere, iterate, pivot and stop, with evidence
  • Know how the week runs, with your tools set up for Day 2

OUTPUT · Set-up checklist complete and a one-line idea in product/01-route.md