Planning Before Building
Most project disasters have one cause: someone started building before they knew what they were actually building. They began coding, prompting an AI, or designing screens without answering the most basic questions: Who is this for? What does it do? What does done look like? Planning is not bureaucracy. A good plan takes fifteen minutes and saves you hours of undoing the wrong work.
The Four Planning Questions
Before writing a single line of code or sending a single AI prompt, answer these four questions in writing: 1. Who is the user? Name a real type of person — not "everyone." "A sixth grader studying for a science test" is useful. "People who like apps" is not. 2. What is the one core thing this does? Not ten features — one. If you can only ship one thing, what is it? 3. What does the finished product look like? Describe the simplest version that would satisfy a real user. This is your Minimum Viable Product, or MVP. 4. What could go wrong? Name one technical risk and one scope risk. A scope risk is when the project keeps growing bigger and bigger until it never ships.
MVP stands for Minimum Viable Product. It is the simplest version of your project that still delivers real value to a real user. An MVP has one feature done well — not ten features done poorly.
Here is a real planning session. A student wants to build a study-help bot using an AI API. Who is the user? Middle schoolers studying vocabulary for English class. One core thing? You type a word; the bot gives you a definition and uses it in a sentence. MVP? A single text box, one button, one output area. No accounts, no history, no fancy design. Risks? Technical: the AI sometimes gives wrong definitions (solution: add a disclaimer). Scope: adding a quiz mode, a leaderboard, a favorites list — none of that is the MVP. With those four answers written down, building becomes straightforward. Every decision has a reference point: does this serve the MVP, or not?
How AI Changes Planning
AI tools can make you feel like planning is unnecessary. You type a vague idea, the AI generates something, and suddenly there is stuff on the screen. It feels like progress. But AI-generated output without a plan usually fails the Shipping Test from Lesson 1. It produces something that looks like a product but does not actually solve the user's real problem. The smarter use of AI is to bring it into the planning phase. You can describe your user, your core feature, and your scope risk — and ask the AI to stress-test your plan. It will spot gaps you missed. Then you build with clarity.
Scope creep means your project keeps growing. You finish the core feature and think, "I should add a leaderboard." Then, "I should add user accounts." Each addition delays shipping. Guard your MVP fiercely.
Prompt Challenge
Write a planning prompt that asks an AI to review your project plan and find weaknesses before you build.
Your prompt should…
- describe the target user specifically by age and goal
- state the single core feature of the MVP clearly
- ask the AI to identify the biggest scope creep risk
What does MVP stand for, and why is it important?
Why is 'everyone' a bad answer to the question 'Who is your user?'
Four-Question Plan
- Step 1: Choose one idea for a simple AI-powered tool or app you would like to build.
- Step 2: Write your answers to the four planning questions: Who is the user? What is the one core thing? What does the MVP look like? What could go wrong?
- Step 3: Circle your scope risk. Underline any feature you wrote down that is NOT in the MVP — those are off-limits until after shipping.
- Step 4: Show your plan to a classmate. Ask them: if you built exactly what I described, would it be worth using? Listen carefully to their answer.