MODULE 2 — PROMPT ENGINEERING MASTERY

6

2.1 What is Prompt Engineering?

Prompt Engineering is the process of designing effective instructions for AI systems to produce useful, accurate and relevant outputs.

A strong prompt generally provides:

Clarity + Context + Format + Constraints + Iteration + Precision

2.2 What is a Prompt?

A prompt is an instruction, question, command or information provided to an AI system to generate a response.

Weak Prompt

"Explain Artificial Intelligence."

This does not specify:

• Audience
• Length
• Depth
• Format
• Purpose

Better Prompt

"Explain Artificial Intelligence in 150 words using three real-world examples for college students."

The second prompt gives the AI a clearer target.

2.3 Four Core Components of a Prompt

5

Task

What should the AI do?

Example:
"Summarise this article."

Context

What background information does the AI need?

Example:
"The summary is for Class 10 students."

Format

How should the response look?

Example:
"Use five bullet points."

Constraints

What limitations should the AI follow?

Example:
"Keep the response under 100 words."

2.4 Types of Prompts

5

Instruction-Based Prompt

Directly tells the AI what to do.

"Write an essay about climate change."

Question-Based Prompt

Asks a focused question.

"What are the benefits of AI in education?"

Contextual Prompt

Provides background before asking the question.

"I am preparing for an AI examination. Explain machine learning in simple language."

Conversational Prompt

Designed for back-and-forth interaction.

"Act as an AI tutor and teach me neural networks."

Creative Prompt

Requests imaginative content.

"Write a short story about a robot helping students learn."

2.5 Zero-Shot, One-Shot & Few-Shot Prompting

5

Zero-Shot Prompting

The AI receives the task without examples.

Example:

"Translate 'Good Morning' into French."

Advantage

Simple and fast.

Limitation

May be less effective for complex tasks.

One-Shot Prompting

The AI receives one example before performing the task.

Example:

Happy = Joyful
Sad = ?

The example establishes the expected pattern.

Few-Shot Prompting

The AI receives several examples.

Example:

Dog = Animal
Rose = Flower
Mango = Fruit
Sparrow = ?

Expected output:

Bird

Few-shot prompting can improve pattern recognition and reduce ambiguity, especially for more complex tasks.

2.6 Role-Based Prompting

5

Role-based prompting tells the AI to respond from a particular professional or expert perspective.

Example

"Act as an experienced mathematics teacher and explain exponents to Grade 8 students."

Possible roles include:

• Teacher
• Data analyst
• Software engineer
• Career counsellor
• Content writer
• Research assistant

Role instructions can help establish the desired perspective, vocabulary and level of detail.

2.7 Step-by-Step Prompting

For complex mathematical or logical tasks, prompts can ask the model to work through the problem systematically.

Example

"Solve this problem step by step: A train travels at 60 km/hr. How far will it travel in 5 hours?"

Calculation:

Distance = Speed × Time

60 × 5 = 300 km

The key idea is to make the task structure explicit and encourage systematic problem solving.

2.8 Prompt Refinement

5

A first prompt does not always produce the ideal response.

Use an iterative process:

Write → Test → Evaluate → Refine → Test Again

Example

Version 1:

"Write about AI."

Version 2:

"Explain AI for beginners."

Version 3:

"Explain AI in 200 words for first-year college students, using three real-world examples and simple language."

Each refinement adds more control.

2.9 Ten Best Practices for Prompt Engineering

1. Define the objective clearly

Know what result you need.

2. Provide sufficient context

Give relevant background information.

3. Use simple and precise language

Avoid ambiguous instructions.

4. Specify output format

Request:

• Bullets
• Tables
• Paragraphs
• JSON
• Code

5. Set constraints

Specify:

• Length
• Tone
• Reading level
• Scope

6. Use examples

Examples help clarify complex patterns.

7. Refine iteratively

Test and improve prompts.

8. Validate AI responses

Verify important facts and recommendations.

9. Maintain ethical standards

Do not use prompts for harmful or misleading purposes.

10. Document successful prompts

Maintain a reusable prompt library.

2.10 Ethical Prompt Engineering

6

Responsible prompting means:

Verify Important Information

Do not blindly trust AI-generated facts.

Protect Privacy

Do not unnecessarily provide sensitive or confidential information.

Avoid Harmful Content

Do not deliberately create misinformation, harmful content or deceptive material.

Respect Intellectual Property

Give credit where required and respect copyright.

Follow Policies

Follow organisational rules and applicable regulations.

Maintain Human Oversight

AI should augment human capabilities, not replace ethical judgement.

Module 2 Practical Activity

Prompt Improvement Challenge

Start with:

"Write a report about AI."

Improve it through three iterations.

Version 1

Add the audience.

Version 2

Add the purpose and format.

Version 3

Add length, tone, examples and constraints.

Then compare all three outputs.