Pattern: Plan First, Then Code

Have the AI write a plan before it starts coding — it prevents expensive wrong turns.

The problem with "just do it"

If you tell an agent "build this" directly, it starts acting immediately — and often makes silent assumptions that turn out wrong. You usually only notice once something's already half-built the wrong way.

The pattern

Ask the AI for a plan before any code gets written:

Describe the goal and requirements
Ask the AI to produce a step-by-step plan (no code yet)
Read the plan, correct wrong assumptions, answer open questions
Only then: "Now implement the plan"

Why this works better

A plan is cheap to read and correct — code is more expensive to read and correct. Catching a mistake in the plan ("you forgot we already have auth middleware") takes a minute. Catching the same mistake in finished code takes much longer.

When it matters most

Multi-step or unclear tasks
When multiple files or systems are involved
When mistakes would be costly (production code, database migrations)

For small, unambiguous changes ("change this color"), an explicit planning step is usually overkill.

What it looks like

You: Before you write any code, give me a plan for adding rate limiting to the API.
Agent: Plan:
1. Add a rate-limit middleware in app.js using a sliding-window counter
2. Apply it to all /api routes, return 429 on limit exceeded
3. Add config for limits per route in config.js
You: Looks good. Just do step 1 for now, I'll review before we continue.
Agent: Starting with step 1 — the middleware in app.js.

EXAMPLE

Prompt: 'First create a plan (no code changes) for a login feature with email+password. I'll confirm the plan before you start.'

🛠️ EXERCISE — TRY IT YOURSELF

Take a small real feature from your project (or: “Build a function that converts CSV rows to JSON”) and make the AI plan first, build second.

Write: “First create a plan in 3–5 steps. No code yet.”
Read the plan critically. Cut one step that is too big, or make it more precise.
Only then: “Implement step 1.” — and work through the plan step by step.

✅ SELF-CHECK

☐ Did the AI really write NO code in the first step? (If it did: sharpen your prompt)
☐ Did at least one plan step need correcting? (Almost always — that is exactly what the plan is for)
☐ Did the result come together in smaller, checkable pieces than usual?

QUICK QUIZ

Why is a planning step before coding worthwhile?