10x Better AI Results with Prompt Engineering - How to Ask ChatGPT the Right Questions

Hello!

Ever asked ChatGPT something and got a completely wrong answer?

If you thought "AI isn't that great..." and gave up, wait!

Actually, how you ask the question completely changes the answer quality.

Today let's talk about prompts that can make you 10x better at using AI!

What is a Prompt?

💡 Definition of Prompt

A Prompt is a command you give to AI.

Simply put,

• What you type into Google search is a "search term"
• What you type into an AI chatbot is a "prompt"

But it's not just throwing keywords,

It's key to specifically request "how you'd like it to respond"!

Bad Prompts vs Good Prompts

Bad Prompt Example

Too vague, right?

AI doesn't know what you want, so it gives an ambiguous answer.

The response feels like a textbook and isn't very helpful.

Good Prompt Example

Much more specific, with actionable advice you can implement right away!

4 Principles of Good Prompts

1. Assign a Role

Give AI a role like "You are a ○○ expert!"

The more clear the role, the more the response tone and style changes.

2. Be Specific

Specific questions get much better answers than vague ones.

Bad example:

Tell me about exercise

Good example:

Tell me 5 beginner bodyweight exercises I can do at home for 30 minutes. Include sets and reps for each exercise.

The more specific, the more AI gives you exactly what you need!

3. Specify Format

Tell it "what format you'd like the response in."

Examples:

Organize the differences between ChatGPT and Claude in a table

Explain Python beginner learning order divided into stages 1 through 5

Write an email in a polite tone, under 200 characters

Setting the format in advance gets you responses in your desired style.

4. Provide Examples

Show AI the style you want with examples, and it'll follow!

Example:

Explain in a simple tone like this example:

Example: "AI is technology that learns from data and finds patterns. Like how humans learn through experience!"

Topic: What is machine learning?

This way AI follows the example tone.

Practical Prompt Templates

Here are some templates you can copy and use right away!

Template 1: Learning Helper

You are a [topic] expert instructor.
Explain [concept] so [audience (e.g., elementary students, college students)] can understand
using easy examples in 3 paragraphs.

Real usage:

You are an economics expert instructor.
Explain "inflation" so middle schoolers can understand
using easy examples in 3 paragraphs.

Template 2: Writing Assistant

Write a blog post about [topic].
Tone: [casual/polite/professional]
Length: [how many paragraphs or characters]
Content to include: [specific requirements]

Real usage:

Write a blog post about "How to Use AI".
Tone: Casual and friendly
Length: 5 paragraphs (3-4 sentences each)
Content to include: 3 ChatGPT usage examples, beginner tips

Template 3: Code Writing Helper

Write code in [language] that implements [function].
Conditions:
- [condition 1]
- [condition 2]
Add comments to each line.

Real usage:

Write Python code that reads an Excel file and extracts specific columns.
Conditions:
- Use pandas library
- Extract only "Name" and "Age" columns
Add comments to each line.

Template 4: Idea Brainstorming

Suggest [number] ideas about [topic].
For each idea include:
- Title
- One-line description
- Pros and cons

Real usage:

Suggest 5 ideas about "improving remote work productivity".
For each idea include:
- Title
- One-line description
- Pros and cons

Prompt Upgrades: Before & After

The same question gets completely different answers depending on how you ask!

Case 1: Travel Planning

Before (vague question):

Recommend Japan travel

AI Response: Tokyo, Osaka, Kyoto are good. Each city has many famous tourist spots... (too general)

After (specific question):

I'm planning a 3-night 4-day Japan trip.
- Travel period: April (cherry blossom season)
- Budget: 1.5 million won per person
- Interests: food tours, traditional culture experiences
- Departure: Seoul

Create a day-by-day itinerary. Include estimated costs and travel time for each schedule.

AI Response: Much more detailed itinerary + transportation + restaurant recommendations + budget allocation!

Case 2: Resume Writing

Before:

Write a resume

AI Response: Only provides generic resume template...

After:

I'm writing a resume for an entry-level marketing position.

Background:
- Education: Business major graduate
- Experience: 6 months SNS marketing intern
- Main achievement: Grew Instagram followers from 500 to 5,000
- Target company: Startup (B2C mobile app)

Write a cover letter draft in 3 paragraphs:
Paragraph 1: Motivation
Paragraph 2: Key experiences and achievements
Paragraph 3: Aspirations

AI Response: Customized cover letter complete!

Case 3: Coding Question

Before:

How to solve Python error

AI Response: Lists general error types...

After:

I got this error when running Python code:

[Error message]
TypeError: unsupported operand type(s) for +: 'int' and 'str'

Code:
age = 25
message = "Age is " + age

Explain why the error occurs and show me the corrected code.

AI Response: Exact cause + solution + corrected code provided!

Common Prompt Writing Mistakes

Mistake 1: Questions Too Short

Bad example:

Recommend blog topics

AI doesn't know what to recommend. What field? Target audience? Purpose?

Good example:

Recommend 10 technical blog topics for IT professionals.
Focus on AI, cloud, and web development fields.

Mistake 2: Requesting Too Much at Once

Bad example:

Tell me how to learn Python, recommend books, and online courses. Also give me 10 project ideas.

Too many questions make AI superficially answer each item.

Good example:

Organize a Python beginner learning roadmap in 3 stages.
Include learning period and main content for each stage.

Ask one thing at a time, then follow up if needed!

Mistake 3: No Context Provided

Bad example:

Fix this:
"When will you send the report?"

AI doesn't know who it's being sent to or what tone to use.

Good example:

This is an email to my boss, make it more polite:
"When will you send the report?"

Advanced Prompt Techniques

1. Chain of Thought (Thinking Step-by-Step)

A technique to make AI think step-by-step.

Regular question:

What is 257 × 34?

Chain of Thought question:

Calculate 257 × 34.
Show the step-by-step solution process.

This makes AI answer more accurately!

2. Few-shot Learning (Example Learning)

Show examples of the style you want.

Example:

Translate English sentences to Korean like this:

Input: "How are you?"
Output: "How's it going?"

Input: "See you later"
Output: "Catch you later"

Input: "I'm on my way"
Output:

AI will translate following the example style!

3. Role Playing (Acting Out Roles)

Assign AI a specific role.

Example:

You are a career coach with 10 years of experience.
A job seeker shared their concern: "I get nervous during interviews."
Empathize and give 3 specific pieces of advice.

The clearer the role, the higher the response quality!

Practice Exercises

Try improving prompts yourself!

Problem 1

Bad prompt:

Tell me about exercise

Improvement direction: Make it specific (goal, time, location, level, etc.)

Tell me a beginner full-body workout routine I can do at home for 30 minutes.
Recommend 5 bodyweight exercises without equipment.
Include sets, reps, and precautions for each exercise.

Problem 2

Bad prompt:

Cooking recipe

Improvement direction: What dish? How many servings? Difficulty? Time?

Tell me a pasta recipe for 2 servings that beginners can make in 30 minutes.
Include ingredient list, cooking steps, and cooking tips, explained step-by-step.

Prompt Checklist

Check this checklist before asking!

• Did you assign a role? (e.g., You are a ○○ expert)
• Did you provide specific context? (e.g., for beginners, within 30 minutes)
• Did you specify the format? (e.g., as a table, in 3 paragraphs)
• Did you provide examples? (when needed)
• Did you ask about only one topic at a time?

If 3 or more out of 5 are checked, it's a good prompt!

Wrapping Up

Just writing prompts well can increase your AI utilization 10-fold!

It might feel awkward at first, but you'll get used to it quickly with practice.

In the next article, aren't you curious about "how AI actually learns?"

I'll explain the training process simply—how it consumes data and grows!

Next Article Preview: 📌 How is AI Trained? The Secret of AI Learning from Data