KOSUNI KOREA
Korean News in English
Technology

Mastering AI Prompt Writing: The Key to Getting Better Results from AI Tools

Published October 16, 2026 · 1 views
Mastering AI Prompt Writing: The Key to Getting Better Results from AI Tools

Why Are Prompts Important?

Many people using AI often say:

'The AI made something weird.'

But in most cases, it's not the AI's fault—it's because the prompt was too vague.

For example, if you go to a restaurant and say:

'Give me delicious food.'

What will you get? It could be Korean cuisine, Chinese food, or something spicy.

But if you say:

'I can't eat spicy food, so please give me a non-spicy pork cutlet meal.'

You're much more likely to get what you want.

The same principle applies to AI.

Since AI generates results based on your input, the more specific your prompt, the better the outcome.


What Is a Prompt?

A prompt (or 'prompt') is essentially an instruction given to AI.

Everything you type into AI counts as a prompt.

For example:

'Create a login screen.'

This is also a prompt.

But a good prompt isn't just a simple request—it includes all the necessary details to get the desired result.


Examples of Poor Prompts

The following requests make it difficult for AI to understand your intent:

Create a login screen.

Make it look nice.

Make it in the latest style.

Add member functionality.

These prompts are too abstract.

Just like people have different ideas of what 'nice design' means, AI can't know exactly what you want without more details.


Examples of Good Prompts

Here's how to make your prompts a bit more specific:

Create an email login screen using Next.js App Router and Tailwind CSS.

Support mobile responsiveness,
Make the login button blue,
And include a sign-up button as well.

By including:

  • The technology you're using

  • Your desired features

  • Your design preferences

  • Your constraints or conditions

AI can generate much more accurate results.


The Basic Structure of a Good Prompt

A good prompt typically follows this order:

  1. Role

  2. Technology

  3. Task Description

  4. Conditions

  5. Desired Output Format

If you remember these five elements, you'll get great results in most situations.


Start by Defining the Role

Telling AI what role to play helps it provide more relevant responses.

For example:

You are a Next.js expert.

Or:

You are a React developer.

Or:

You are a UI designer.

By defining the role first, AI will respond from that perspective.


Specify the Technology You're Using

AI knows many technologies, so it's important to specify which ones you want to use.

For example:

Next.js App Router
TypeScript
Tailwind CSS
Supabase

By listing the technologies, you can avoid unnecessary code.


Describe Desired Features Clearly

Instead of:

Create a sign-up feature.

Write something like this:

Create an email registration feature.

Password must be at least 8 characters,
Include email verification as well.

This gives AI a clearer picture of what you need.


Include All Necessary Conditions

The more conditions you provide, the better AI can deliver accurate results.

For example:

  • Mobile responsive design

  • Dark mode support

  • Use TypeScript

  • Use Tailwind CSS

  • Component separation

  • Add comments to code

You can include any relevant conditions like these.


Specify the Desired Output Format

You can also tell AI how you want it to respond.

For example:

Just provide the code.

Or:

Explain it in a way that beginners can understand.

Or:

Break it down step by step.

This helps you get the exact format you need.


Example Prompts for Real Projects

Now, let's look at some prompts we'll use to build a dating app:

Example 1

You are a Next.js expert.

Create the login screen for a dating app using Tailwind CSS.

Email input,
Password input,
Login button,
Sign-up button.

Make it look good on mobile devices too.


Example 2

Create a users table using Supabase Database.

Columns:
id,
email,
nickname,
created_at

Also provide the SQL code.


Example 3

Create a profile card UI.

Include:
Profile photo,
Nickname,
Age,
Location,
Like button.

Make it mobile responsive.


What If AI Doesn't Produce the Desired Result?

You won't get perfect results on the first try.

Instead of creating a new prompt, give specific feedback to improve the existing result.

For example:

The buttons are too big.

Show it in one line on mobile devices.

Change the color to shades of purple.

The code is too complex.

Separate into components.

By giving specific feedback, AI will provide a revised version that better matches your needs.


Vibe Coding Emphasizes Conversation

Many beginners ask once and give up if the result isn't perfect.

But Vibe Coding is about continuing the conversation to reach your goal.

For example:

  1. Create a login screen.

  2. Make the buttons larger.

  3. Adjust for mobile responsiveness.

  4. Add dark mode support too.

This iterative approach is key to getting exactly what you want.


The Prompt Writing Principles You'll Use Going Forward

Remember these five steps for all future prompts:

  • Define the role

  • List the technologies used

  • Explain desired features

  • Add conditions

  • Specify output format

If you follow these five steps, your AI results will improve significantly.


Next Lecture Preview

In the next lesson, we'll cover Git and GitHub from the basics—explaining concepts, creating repositories, and understanding version control principles. We'll also learn how to connect with AI-powered GitHub MCP to let AI commit changes and manage projects directly.


Conclusion

Vibe Coding isn't about giving commands to AI—it's a collaborative process of building your project together with AI.

Good prompts lead to better code, which improves the overall quality of your projects.

The prompt writing skills you've learned today will be essential not just for this dating app project but for any future AI development tools you use.

In our next lesson, we'll dive into Git and GitHub to get closer to real-world development environments.

FAQ

Why is it important to specify the role in an AI prompt?

Specifying a role helps AI provide more relevant responses by understanding the perspective from which you want the output.

What makes a good AI prompt better than a vague one?

A good prompt includes specific details about technology, features, design preferences, conditions, and desired output format, leading to more accurate results.

Korean Source

This article is an English localization of a Korean-language report from BlogScore. Original: [바이브코딩 강좌] 5강. AI에게 일을 잘 시키는 프롬프트 작성법