The Art of Prompt Engineering

1. Introduction & Context

1.1. Overview
Objective:

This lecture introduces prompt engineering—a skill essential for optimizing AI

model interactions. Attendees learn how to craft, refine, and optimize prompts
to achieve specific, high-quality outputs from AI systems, which is highlighted as
one of the highest-paying roles in AI today.
2. Main Content

2.1. Introduction to Prompt Engineering

2.1.1. Significance of Prompt Engineering

Key Points: Described as an “amazing and mind-blowing skill.” Recognized by
the World Economic Forum as the highest paying AI job. The goal is to optimize
AI interactions for maximum efficiency and accuracy.

2.2. Understanding Prompt Engineering

2.2.1. What Are Prompts?

Definition: Prompts are the instructions, context, or input provided to AI
models to perform specific tasks.
Purpose: They guide AI models to generate relevant and accurate responses.

2.2.2. Prompt Engineering Defined

Definition: The art of crafting and optimizing prompts to achieve the best
possible outputs from AI.
Key Challenge: Addressing issues where AI misinterprets vague instructions
and delivers unexpected results.
Objective: To learn how to write detailed, specific, and context-rich prompts.
2.3. Agenda for the Session
Introduction to Prompt Engineering
Why Prompt Engineering Matters for Professionals
Key Concepts in Prompt Engineering
Designing Prompts for Real-World Tasks
Advanced Techniques for Better Results
Practical Tips and Strategies for Generating Good Prompts
Tools and Templates for Generating Prompts
Outcome: By the end of the session, attendees will be able to create optimized
prompts for any available Large Language Model (LLM).

2.4. Goal for Today

Primary Goals: Learn structured, optimized, and advanced prompting
techniques. Understand that well-structured prompts improve AI accuracy and
relevance.

2.5. Examples: Bad Prompt vs. Good Prompt

Bad Prompt:

"Explain marketing strategy"

Issue: Too vague, lacks specifics.

Good Prompt:

"You are a marketing consultant; provide 5 marketing strategies for small

businesses, emphasizing low-cost digital marketing methods."

Benefit: Specific context and clear instructions lead to better, targeted outputs.
2.6. Types of Prompt Engineering

2.6.1. Context and Constraint-Based Prompting

Definition: Using specific background information (context) and clear
boundaries (constraints) to guide AI.
Importance: Reduces ambiguity and tailors outputs to user needs.
Examples:
Product Description: Weak Prompt: "Write a product description for a
smartphone." Improved Prompt: "Write a product description for the XYZ pro
12 smartphone highlighting its 108-megapixel camera, 120 hertz OLED display,
and 5,000 mAh battery. The tone should be engaging and appealing to tech
enthusiasts."
Translation: Weak Prompt: "Translate the sentence." Strong Prompt: "Translate
the following sentence from English to Hindi, ensuring the cultural nuances are
preserved: 'The Festival of Lights brings joys to everyone.'"
News Article: Weak Prompt: "Write a news article." Strong Prompt: "Write a 300-
word news article suitable for the Times of India about the recent
advancements in AI technology in Bangalore, focusing on local startups."

Real-World Applications: Customer support emails, educational content, and

healthcare communication.
2.6.2. Role-Based Prompting
Definition: Assigning the AI a specific role to tailor its responses.
How It Works: Simulates expertise, adjusts tone and style, and enhances
engagement.
Examples:
Interview Questions: Without Role-Based Prompting: "Give me answers for a
few common interview questions." With Role-Based Prompting: "As an
experienced HR professional, craft well-structured answers for common
interview questions. Make them confident, concise, and aligned with best hiring
practices."
Resume Writing: Weak: "Help me write a resume." Strong: "As a professional
career coach, craft a compelling resume for a software engineer with 5 years of
experience in Python and AI, ensuring it aligns with current hiring trends."
Investment Advice: Combining role-based and context/constraint-based
prompting for tailored advice.

2.6.3. Zero-Shot and Few-Shot Prompting

Definition: Zero-Shot: Instructing AI without examples. Few-Shot: Providing a
few examples before asking the task.
Why It Works: Few-shot prompting helps AI recognize patterns, ensuring
responses align with expectations.
Examples:
Product Descriptions: Zero-Shot: "Write a product description for a
smartphone." Few-Shot: Provide two examples of premium smartphone
descriptions, then ask for a tailored description.
Social Media Posts: Provide examples of posts before asking AI to generate a
new one in a similar style.

2.6.4. Advanced Prompting Techniques

2.6.4.1. Chain of Thought (CoT) Prompting

Definition: Instructing AI to explain its reasoning step by step before providing
the final answer.
Examples:
Customer Complaint Response: General: "Write a response to a customer
complaining about a delayed order." CoT: "Generate a professional and
empathetic response to a customer frustrated about a delayed order. Structure
it step by step: 1. Acknowledge the customer's frustration; 2. Apologize; 3.
Explain the cause of the delay; 4. Suggest next steps."
Investment Advice Analysis: "Analyze if a middle-class investor in India should
allocate 10 lakhs in mutual funds. Break it down step by step."

2.6.4.2. Tree of Thought (ToT) Prompting

Definition: Encourages AI to explore multiple reasoning pathways before
arriving at a final answer.
Example: "Analyze the pros and cons of three possible market entry strategies
for international expansion: Joint Venture, Direct Investment, and Franchising.
Recommend the best strategy based on scalability and risk."
Benefit: Provides comprehensive analysis and aids in decision-making where
multiple solutions exist.

2.6.4.3. Self-Consistency Prompting

Definition: Requesting multiple independent responses to the same prompt
and selecting the most consistent answer.
Examples: Generating several sets of employer retention strategies or salary
negotiation tactics, then summarizing common effective strategies.
Benefit: Reduces errors by considering multiple viewpoints.

2.6.4.4. ReAct Prompting (Reason + Act + Reflection)

Definition: Involves prompting AI to reason through a problem, act by

providing a step-by-step answer, and reflect to ensure completeness.
Examples: Business Report Summarisation—"Analyze this report using the
following steps: 1. Identify key trends; 2. Extract funding patterns; 3. Verify all
critical insights are covered; 4. Summarize the findings."
Benefit: Ensures logical and structured responses with fewer omissions.

2.6.4.5. Multimodal Prompting

Definition: Combining multiple input formats (text, images, audio, etc.) in one
prompt.
Examples: "Upload an image of the Taj Mahal and prompt: 'Analyze this image
and provide a historical summary.'"
Real-World Applications: Marketing (analyzing ad images), education
(explaining diagrams), UI/UX design, and data analysis.

2.7. Real-World Applications and Use Cases

2.7.1. Industry-Specific Examples

Product Requirement Analysis (Software Development): Role-based

prompting with Chain of Thought to break down features.
Competitive Market Analysis (Marketing & Strategy): Self-Consistency
prompting for generating multiple analyses.
Customer Support: Role-based plus context & constraint-based prompting to
draft precise responses.
Investment Portfolio Strategy (Finance): Using Chain of Thought prompting for
detailed step-by-step analysis.
Product Launch Content (Marketing): Few-Shot prompting combined with
role-based prompting.
Bug Troubleshooting (IT Support): ReAct prompting for structured problem
solving.
2.7.2. Resources for Learners
Prompt Engineering Handbook: A summary of session content, best practices,
and prompting techniques (Access Here).
Ready-to-Use Prompt Formulas: Pre-made templates for various tasks
(content writing, HR, emails, etc.) (Download Templates).
Prompt Generator (Jarvis): An interactive tool to generate advanced-level
prompts through clarifying questions and tailored outputs (Try It Now).

3. Key Insights & Takeaways

Fundamental Understanding: Prompt Engineering is an essential skill in
optimizing AI interactions for quality and efficiency.
Types & Techniques: Different prompting strategies (context-based, role-
based, zero/few-shot, CoT, ToT, Self-Consistency, ReAct, and multimodal) cater
to various needs and outputs.
Practical Impact: Well-defined and optimized prompts can transform generic AI
responses into highly tailored outputs.
Real-World Applications: The techniques apply across industries—from
customer support to investment advice—demonstrating versatility.
Resources: Learners gain exclusive access to handbooks, prompt templates,
and prompt generators that further enhance their prompt engineering skills.
4. Conclusion & Summary

4.1. Summary of Core Message

Core Message: The session provided a comprehensive overview of prompt
engineering techniques. By exploring various methods—from context-based
and role-based prompting to advanced strategies like Chain of Thought and
multimodal prompting—attendees learn how to refine AI interactions for better
accuracy, relevance, and efficiency.

4.2. Next Steps & Further Reading

Next Steps: Practice implementing these prompting techniques with real-world
tasks. Experiment with the provided resources (handbook, prompt formulas,
prompt generator) to master prompt engineering.
Further Reading: Additional materials on AI prompting strategies. Relevant
articles and case studies on advanced prompt engineering in professional
settings.

5. Additional Enhancements for Readability &

Engagement

5.1. Comparison Table: Prompting Techniques

Technique Definition Key Benefit

Context & Provides detailed Reduces ambiguity and

Constraint- background and sets clear tailors output to specific
Based boundaries needs

Role-Based Assigns a specific role to Adjusts tone, expertise, and

Prompting the AI style for more professional
 Technique Definition Key Benefit
output

Instructs with no or few Helps AI recognize patterns

Zero-Shot/Few-
examples to guide and align output with
Shot
response format expectations

Requires step-by-step
Chain of Enhances logical reasoning
reasoning before the final
Thought (CoT) and clarity in complex tasks
answer

Explores multiple
Tree of Thought Useful for decision-making
reasoning paths before
(ToT) in complex scenarios
finalizing an answer

Generates multiple Reduces errors by ensuring

Self-
responses to select the a consensus across
Consistency
most consistent answer responses

Involves reasoning, acting,

ReAct Ensures comprehensive
and reflecting in a
Prompting and well-verified output
structured process

Uses multiple input types Enhances understanding

Multimodal
(text, image, etc.) for through diverse data
Prompting
richer responses formats

5.2. Self-Check Q&A

Q: What is prompt engineering and why is it important?
Q: How does role-based prompting improve AI responses?
Q: What is the key difference between zero-shot and few-shot prompting?