Iteration

The first answer from AI is a draft, not the final result. Iteration is how you turn that draft into exactly what you want.

What is iteration?

Think about getting a haircut. The barber cuts, you look in the mirror, and you say "a little shorter on the sides". You don't start over. You adjust.

Iteration means improving something in small rounds. With AI, each round looks like this:

• Read the answer carefully.
• Decide what is good and what is missing.
• Ask for one clear change.
• Check the new answer and repeat.

Iteration in action

Let's improve a welcome message for the Habit Tracker app over three rounds.

Round 1: the first draft

Write a welcome message for the home screen of my habit tracker app.

The answer is fine, but it's 4 sentences long and a bit formal. That is normal. Now we refine.

Round 2: fix the length

Good start. Make it shorter: one sentence, under 15 words.

Round 3: fix the tone

Perfect length. Now make it warmer and more casual, like a friend cheering you on.
Keep everything else the same.

Let's break this down:

• Each follow-up asks for one change, so you can see what that change did.
• "Good start" and "Perfect length" tell the AI what to keep.
• "Keep everything else the same" stops the AI from rewriting parts you already liked.

Useful follow-up phrases

You'll use these again and again. Copy the ones you like.

Make it shorter / longer.
Use simpler words, for a complete beginner.
Keep the structure, but change the tone to [friendly / professional].
Option 2 is the best. Give me 3 more like option 2.
That part is wrong: [explain]. Fix only that part.
Keep everything else exactly the same.
Before changing anything, tell me what you plan to change.

When to start a new chat

Sometimes a chat goes off track. The AI keeps repeating a mistake or mixes up old instructions. When that happens, a fresh start is faster.

• Ask the AI: "Summarize what we've decided so far in a short list."
• Open a new chat.
• Paste the summary plus your improved prompt.

This works because long chats fill up the context window. Older details get less attention, so a clean summary helps the AI focus.

Save your good prompts

When a prompt works really well, save it. A prompt library is just a note or document where you keep prompts you can reuse.

## Short app welcome message
You are a friendly copywriter. Write a welcome message for the home screen
of my [type of app]. One sentence, under 15 words, warm and casual.

## Explain a concept simply
You are a patient teacher. Explain [concept] to someone who has never coded.
Use one everyday analogy and keep it under 100 words.

Common mistakes

• Giving up after the first answer, or accepting a "meh" result.
• Changing five things in one follow-up, so you can't tell which change helped.
• Saying "no, that's wrong" without saying what is wrong.
• Rewriting the whole prompt from scratch when one small tweak would do.
• Losing a great prompt because you didn't save it.

What's next?

Sometimes describing what you want takes many rounds. A faster way is to show the AI an example. Let's look at that next.