🎯 Prompt Engineering: Instruction Design

The Recipe Analogy

Imagine you’re giving someone a recipe to bake a cake. If you just say “make a cake,” they might be confused. But if you say “take 2 cups of flour, add 3 eggs, mix slowly, then bake at 350°F for 30 minutes” — now they can actually make something delicious!

Talking to AI is exactly like giving a recipe. The clearer your instructions, the better the result!

1. Instruction Clarity

What Is It?

Instruction clarity means saying exactly what you want in a way that’s impossible to misunderstand.

Think of it like this: If you tell your friend “get me something to drink,” they might bring water, juice, or soda. But if you say “please bring me a cold glass of orange juice” — you get exactly what you want!

Why It Matters

• AI doesn’t guess well
• Clear instructions = Clear results
• Saves time (no back-and-forth)

Simple Example

❌ Unclear: “Write about dogs”

✅ Clear: “Write 3 sentences about why golden retrievers make great family pets”

The Clarity Checklist

✓ What exactly do I want?
✓ How long should it be?
✓ What style or tone?
✓ Any specific format?

2. Specificity in Prompts

What Is It?

Specificity means adding details that narrow down what you want. It’s like zooming in on a map — the more you zoom, the more precise your location.

The Zoom Analogy

🌍 World → 🗺️ Country → 🏙️ City → 🏠 Street → 📍 House

Each step gets MORE specific!

Simple Example

❌ Vague: “Help me with math”

✅ Specific: “Solve this equation: 2x + 5 = 15. Show each step.”

Types of Specificity

3. Providing Context

What Is It?

Context is the background information that helps AI understand your situation. It’s like telling a doctor your symptoms before they prescribe medicine!

The Doctor Analogy

🏥 Bad: “I need medicine” 🏥 Good: “I have a headache, it started this morning, I’m allergic to aspirin”

Why Context Helps

Without context, AI has to guess. With context, AI knows:

• Your situation
• Your constraints
• Your goals

Simple Example

❌ No Context: “Write an email”

✅ With Context: “I’m a teacher writing to a parent. Their child missed 3 days of school. I need to ask if everything is okay and offer makeup work. Keep it friendly but professional.”

Context Building Blocks

graph TD
    A[WHO are you?] --> E[Complete Context]
    B[WHAT do you need?] --> E
    C[WHY do you need it?] --> E
    D[WHO is it for?] --> E

4. Prompt Priming

What Is It?

Prompt priming is warming up the AI by telling it what role to play or what style to use BEFORE asking your question.

Think of it like an actor getting into character before a scene!

The Actor Analogy

🎭 Before: “Can you explain photosynthesis?”

🎭 After: “You are a fun science teacher who loves using jokes and simple examples. Explain photosynthesis to a curious 8-year-old.”

Simple Example

❌ No Priming: “How do I fix my code?”

✅ With Priming: “Act as a patient coding mentor. I’m a beginner learning Python. My code isn’t working and I’m getting a ‘TypeError’. Help me understand what’s wrong and how to fix it.”

Priming Templates

5. Directional Stimulus

What Is It?

Directional stimulus is pointing the AI in the right direction by giving hints about what kind of answer you want.

It’s like telling a treasure hunter “the gold is somewhere in the forest” instead of “find gold somewhere on Earth”!

The Compass Analogy

🧭 Without direction: AI wanders everywhere 🧭 With direction: AI goes straight to what you need

Simple Example

❌ No Direction: “Tell me about space”

✅ With Direction: “Focus specifically on how astronauts sleep in zero gravity. Include at least 2 surprising facts.”

Directional Keywords

Use these words to steer the AI:

• “Focus on…” - Narrows the topic
• “Emphasize…” - Highlights what matters most
• “Compare…” - Gets contrasts
• “Avoid…” - Steers away from unwanted content
• “Include…” - Ensures certain elements appear

Steering Examples

"Focus on practical tips, not theory"
"Emphasize cost-saving benefits"
"Compare the pros and cons"
"Avoid technical jargon"
"Include real-world examples"

6. Anchoring Techniques

What Is It?

Anchoring means giving the AI a starting point or reference to build from. It’s like showing someone a sample before asking them to create something similar.

The Blueprint Analogy

🏗️ Building without plans = Random results 🏗️ Building with a blueprint = Exactly what you pictured

Types of Anchors

1. Example Anchor - Show what you want
2. Style Anchor - Reference a known style
3. Format Anchor - Provide a template

Simple Example: Example Anchor

❌ No Anchor: “Write a product description”

✅ With Anchor: “Write a product description like this example: ‘The SunBright Lamp brings warm, natural light to any room. Perfect for reading or relaxing. Energy-efficient LED bulbs last 10+ years.’ Now write one for a wireless mouse.”

Simple Example: Style Anchor

✅ “Write in the style of a friendly newsletter” ✅ “Make it sound like a TED talk” ✅ “Match the tone of a children’s book”

Simple Example: Format Anchor

✅ “Use this format: Problem: [describe issue] Solution: [your fix] Result: [what happened]”

7. Scaffolding Prompts

What Is It?

Scaffolding means breaking a big task into smaller steps. Just like building a house floor by floor, you help the AI tackle one piece at a time.

The Staircase Analogy

🪜 Instead of jumping to the top, you climb step by step!

graph TD
    A[Step 1: Gather Info] --> B[Step 2: Organize]
    B --> C[Step 3: Draft]
    C --> D[Step 4: Refine]
    D --> E[Final Result!]

Why Scaffolding Works

• Complex tasks become manageable
• Each step can be checked
• Better quality output
• Easier to fix problems

Simple Example

❌ One Big Ask: “Write a complete business plan for a coffee shop”

✅ Scaffolded Approach:

Step 1: “First, list 5 unique features my coffee shop could have”

Step 2: “Now, for the top 3 features, explain who would be the target customers”

Step 3: “Create a simple pricing menu for 10 items”

Step 4: “Write a one-paragraph summary of the business concept”

Scaffolding Trigger Words

• “First…”
• “Then…”
• “Next…”
• “After that…”
• “Finally…”
• “Let’s start with…”
• “Now let’s move to…”

🎮 Quick Reference: All 7 Techniques

🚀 Putting It All Together

Here’s a prompt that uses ALL 7 techniques:

[PRIMING] You are a friendly fitness coach
who specializes in home workouts.

[CONTEXT] I'm a busy parent with 20 minutes
per day and no gym equipment.

[CLARITY] Create a simple workout routine.

[SPECIFICITY] Include exactly 5 exercises,
each taking about 4 minutes.

[DIRECTION] Focus on full-body movements.
Avoid anything that requires jumping
(apartment living!).

[ANCHORING] Format each exercise like this:
- Exercise Name
- How to do it (2 sentences)
- Muscles worked

[SCAFFOLDING] Let's start with warm-up
exercises first, then move to the main workout.

🌟 You’ve Got This!

Remember: Talking to AI is like giving a recipe.

The more specific, clear, and structured your instructions, the more delicious your results will be!

Start simple, practice often, and watch your prompts get better every day. You’re already on your way to becoming a prompt engineering pro! 🎉