Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Module 2: Prompt
Engineering and Creative
Communication

Unit 1: Introduction to Prompt

Engineering
1.1 What is Prompt Engineering?
Simple Definition:
Prompt engineering is the art and science of writing clear, effective instructions that
guide AI language models (like ChatGPT) to generate the desired output consistently and
accurately.

Formal Definition:
Prompt engineering is the process of crafting and optimizing inputs (prompts) to
language models to produce high-quality, relevant, and contextually appropriate responses
that meet specific requirements.

Key Concept:
Think of a prompt as a question or instruction you give to an AI system. The quality of your
prompt directly affects the quality of the response you receive.

Core Principle:

"The better you communicate your request, the better the AI understands and
responds."

Why Is Prompt Engineering Important?

1. Consistency – Well-written prompts produce consistent, reliable outputs

2. Accuracy – Better prompts lead to more accurate and relevant responses
3. Efficiency – Reduces need for follow-up corrections and iterations
4. Control – Allows steering AI behavior toward your specific goals
5. Cost Reduction – Fewer iterations mean less API usage and lower costs
6. Time Saving – Gets desired output faster without multiple attempts
Real-World Analogy:

 Poor instruction to ChatGPT: "Write something about AI"

Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Better instruction: "Write a 200-word summary of how artificial intelligence is

used in healthcare, explaining three specific applications"
The second prompt is more likely to produce exactly what you need.

1.2 The Evolution of Prompt Engineering

Timeline of Prompt Engineering Development:

Period Development Significance

2018- Introduction of GPT-2 and First large language models (LLMs) that
2020 GPT-3 required prompts to function
Early Discovery of "in-context Researchers found models could learn
2021 learning" from examples in prompts
Mid 2021 Few-shot prompting Showing examples within prompts
emergence improves results dramatically
Late 2021 Chain-of-Thought (CoT) Breaking problems into steps improves
prompting discovered reasoning
2022 ChatGPT launch (November First mainstream AI chatbot;
2022) popularizes prompt engineering
2023 Prompt engineering becomes Organizations hire "prompt engineers";
a skill courses proliferate
2023- Advanced techniques (Self- More sophisticated methods for complex
2024 Consistency, ReAct) reasoning
2024- Multimodal prompting Prompting for images, audio, and video
Present alongside text

Key Milestone: ChatGPT Launch (November 2022)

 Brought AI to mainstream audience

 Made prompt engineering accessible to everyone
 Demonstrated enormous potential of well-designed prompts
 Led to explosion in prompt engineering research and applications
Why Evolution Matters:
Early prompt engineering was mainly trial-and-error. Today, we have scientifically-
validated techniques that work consistently across different AI models.

1.3 Types of Prompts

There are several categories of prompts, each serving different purposes:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Type 1: Direct/Simple Prompts

Definition:
Straightforward questions or instructions with minimal context.

Structure:
"[Task] + [What you want done]"

Examples:

 "What is photosynthesis?"
 "Summarize this article: [article text]"
 "Write a poem about winter"
When to use:

 Simple, factual questions

 General knowledge queries
 Quick tasks
Advantages:

 Easy to write
 Fast response
 Works for straightforward tasks
Limitations:

 May lack specificity

 Can produce generic responses
 Might miss nuance

Type 2: Instructional Prompts

Definition:
Clear, step-by-step instructions that guide the AI through a specific process.

Structure:
"You are [role]. Your task is to [objective]. Follow these steps:

1. [Step 1]
2. [Step 2]
3. [Step 3]"
Example:
"You are a technical writer. Your task is to explain machine learning to a high school student.
Use simple language, include one real-world example, and keep it under 200 words."

When to use:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Complex tasks with multiple steps

 Need specific output format
 Want consistency in responses
Advantages:

 Highly specific results

 Easy to iterate on
 Clear expectations

Type 3: Role-Based Prompts

Definition:
Asking AI to assume a specific persona or role that shapes its perspective and communication
style.

Structure:
"You are [specific role/expert]. [Task]."

Examples:

 "You are a professional copywriter. Write a product description for..."

 "You are a Python expert. Debug this code..."
 "You are a motivational coach. Give advice on..."
When to use:

 When you want specialized knowledge or perspective

 To match a specific communication style
 Professional content creation
Advantages:

 Tailored expertise and tone

 Consistent perspective
 Better domain-specific responses

Type 4: Contextual Prompts

Definition:
Prompts that provide background information and context to guide the AI's understanding.

Structure:
"[Context: situation/background]
[Task: what you want done]
[Requirements: specific constraints]"
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Example:
"Context: I'm a software company with 50 employees. We're expanding to Europe and need to
understand GDPR requirements.
Task: Write an overview of GDPR's impact on our data handling practices.
Requirements: Keep it suitable for non-technical staff, focus on practical implications,
include 2-3 actionable steps."

When to use:

 Complex situations needing background

 Business/professional contexts
 When previous context is important
Advantages:

 Highly relevant responses

 Reduces misunderstandings
 Better aligned with actual needs

Type 5: Open-Ended Prompts

Definition:
Broad, exploratory prompts designed to encourage creative thinking and diverse responses.

Structure:
"[General topic or question]. [Optional: suggest directions to explore]"

Examples:

 "How might AI transform education in the next 10 years?"

 "What are innovative approaches to reducing carbon emissions?"
 "Explore unexpected connections between music and mathematics"
When to use:

 Brainstorming sessions
 Creative thinking exercises
 Generating multiple perspectives
Advantages:

 Encourages creativity
 Multiple valid responses
 Explores possibilities

Type 6: Constrained Prompts

Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Definition:
Prompts with specific constraints (length, format, style, tone) to ensure output meets precise
requirements.

Structure:
"[Task]. [Specific constraints: length, format, tone, etc.]"

Examples:

 "Write a Twitter post (280 characters max) about renewable energy"

 "Create a JSON response with these fields: name, email, status"
 "Write in a formal, professional tone without jargon"
When to use:

 When format matters (social media, code, reports)

 Limited-resource scenarios
 Specific industry/format requirements
Advantages:

 Guarantees specific output format

 Easy to integrate into systems
 Prevents verbose or irrelevant output

1.4 How Does Prompt Engineering Work?

The Basic Cycle:

1. Write/Craft Prompt
↓
2. Submit to AI Model
↓
3. Receive Output
↓
4. Evaluate Quality
↓
5. Is output good? → YES: Use it
→ NO: Refine prompt and repeat
Step-by-Step Explanation:

Step 1: Understand Your Goal

 What exact output do you want?
 Who is the audience?
 What format is needed?
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 What constraints exist?

Example goal: "I need a Python function that validates email addresses"

Step 2: Provide Context

 Share relevant background information
 Explain why this task matters (optional but helpful)
 Mention constraints or requirements
Example context: "I'm building a user registration system for a web application"

Step 3: Give Clear Instructions

 Be specific about what you want
 Specify format, tone, length
 Include examples if possible
 State what to avoid
Example instruction: "Write a Python function that validates email addresses using regular
expressions. Include error handling. Add comments explaining the code."

Step 4: Provide Examples (If Needed)

 For complex tasks, include input/output examples
 Shows the model exactly what you want
 Critical for consistent results

Step 5: Evaluate Output

 Does it meet your requirements?
 Is quality acceptable?
 Are there issues to address?

Step 6: Iterate If Needed

 If output isn't perfect, refine your prompt
 Change wording, add constraints, provide examples
 Re-submit and evaluate
Example Iteration:

Version 1 (Too vague):

"Write something about climate change"
Result: Generic 2-paragraph overview

Version 2 (Better):
"Write a 300-word article about how businesses can reduce their carbon footprint"
Result: Better, but still generic
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Version 3 (Best):
"You are an environmental consultant. Write a 300-word article for small business owners
about practical, cost-effective ways to reduce carbon footprint. Include 3 specific actions with
estimated costs. Keep language non-technical."
Result: Targeted, practical, exactly what needed

1.5 Comprehending Prompt Engineering's

Function in Communication
Prompt Engineering as Communication Bridge:

Prompt engineering is fundamentally about clear communication between human and AI.
It bridges the gap between:

 What humans want

 What AI can provide
 How to make those things match

Communication Levels:
Level 1: Surface Communication

 Simple instruction: "Translate this to Spanish"

 AI translates directly
 No deep understanding needed
Level 2: Contextual Communication

 Includes background: "I'm a tourist in Madrid and need to ask for directions"
 AI understands context
 Tailors response appropriately
Level 3: Strategic Communication

 Includes goal and constraints: "Write marketing copy for tech-savvy millennials that
emphasizes sustainability. Keep it under 100 words and use casual tone"
 AI understands broader intent
 Delivers strategically aligned output

Key Communication Principles:

1. Clarity

 Use simple, unambiguous language

 Avoid vague terms
 Be specific about requirements
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

2. Completeness

 Provide all needed information

 Don't leave important details implicit
 Include examples when helpful
3. Consistency

 Use similar terminology throughout

 Structure prompts logically
 Make expectations obvious
4. Context

 Provide background when relevant

 Explain why the task matters
 Mention constraints upfront
5. Examples

 Show input/output pairs

 Demonstrate desired quality/style
 Make expectations tangible

Communication Challenges and Solutions:

Challenge What Happens Solution
Ambiguous AI interprets literally or Use specific, concrete language
language differently
Missing context Generic or irrelevant Provide background and
response constraints
Conflicting AI confused about priorities Prioritize clearly; use "most
instructions important:"
Unrealistic Disappointing results Understand AI limitations;
expectations iterate
No examples Inconsistent or mediocre Provide 2-3 input/output
output examples

Example of Poor Communication:

Prompt: "Write a blog post about AI"
Problem: Vague topic, unclear audience, no length specified, no style specified
Result: Generic, possibly too long or too short, wrong tone

Example of Effective Communication:

Prompt: "You are a tech journalist. Write a 800-word blog post for Forbes readers about
recent breakthroughs in generative AI. Focus on business applications. Use formal tone.
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Include 3 specific examples of companies using the technology. Start with a compelling
hook."
Why it works: Specific role, clear audience, exact length, specific tone, concrete requirements
(3 examples), structural guidance (starts with hook)
Result: Targeted, professional, exactly what publication needs

1.6 Advantages of Prompt Engineering

Why Companies and Individuals Benefit:

Advantage 1: Improved Efficiency

 How: Well-crafted prompts get desired output in first attempt
 Impact: Saves time iterating or correcting wrong outputs
 Example: A prompt that clearly specifies format means results are immediately
usable in code, documents, or reports
 Benefit: Faster workflows, less manual work

Advantage 2: Higher Quality Output

 How: Specific prompts guide AI toward better responses
 Impact: Output quality matches professional standards
 Example: A role-based prompt that says "You are an award-winning copywriter"
typically generates better marketing copy
 Benefit: Better results without extensive editing

Advantage 3: Cost Reduction

 How: Fewer iterations = fewer API calls (if using paid services)
 Impact: Lower bill and faster processing
 Example: Instead of 5 attempts to get right response, one well-crafted prompt
succeeds
 Benefit: Significant cost savings at scale

Advantage 4: Consistency
 How: Same prompt structure produces similar quality responses
 Impact: Reliable, predictable AI behavior
 Example: A template prompt used across marketing team generates consistent
brand messaging
 Benefit: Standardized outputs; better quality control

Advantage 5: Scalability
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 How: Prompts can be reused and applied to many similar tasks

 Impact: Apply one good prompt to hundreds of documents, queries, or scenarios
 Example: One email-writing prompt can generate dozens of different professional
emails
 Benefit: Apply work once, benefit many times

Advantage 6: Customization and Control

 How: Specific prompts steer AI toward desired behavior
 Impact: Can achieve specific goals, tones, formats
 Example: Same AI model can produce technical writing for experts or simple writing
for children—just change the prompt
 Benefit: Flexibility and control over output

Advantage 7: Reduced Learning Curve

 How: Good prompts act as guides to using AI effectively
 Impact: Even non-experts can get professional-quality results
 Example: A detailed system prompt teaches AI to behave like an expert, so user gets
expert-level output
 Benefit: Democratizes access to AI capabilities

Advantage 8: Better Human-AI Collaboration

 How: Clear prompts bridge the communication gap
 Impact: AI becomes more useful tool for human work
 Example: A well-designed prompt creates a "AI assistant" that understands context
and helps effectively
 Benefit: More productive partnership between human and AI

1.7 The Future of LLM Communication

Emerging Trends:

Trend 1: Multimodal Prompting

 What: Using text, images, audio, and video in prompts
 Current: ChatGPT 4V can process images
 Future: AI will accept mixed-media prompts naturally
 Impact: Richer context; more powerful AI responses
 Example: "Analyze this photo of a broken machine and suggest repairs"
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Trend 2: Agentic AI and Long-Horizon Tasks

 What: AI that can plan and execute complex, multi-step tasks
 Current: AI needs separate prompts for each step
 Future: AI can manage entire workflows with one strategic prompt
 Impact: Autonomous AI assistants that work independently
 Example: "Analyze Q4 sales data, identify trends, create visualizations, and email
report to executive team"

Trend 3: Prompt Optimization and Auto-Prompting

 What: AI that automatically improves prompts or generates optimal prompts
 Current: Humans manually refine prompts
 Future: AI systems that automatically optimize prompts for best results
 Impact: Reduced need for manual prompt engineering
 Example: System says "Your prompt can be improved. Try adding 'specific
examples' for better results"

Trend 4: Personalized Prompting

 What: AI learns and adapts to individual user's preferences
 Current: Same prompt works differently for different users
 Future: AI remembers what works best for each user
 Impact: Increasingly personalized and effective responses
 Example: AI learns "This user prefers formal tone and structured bullet points" and
adapts automatically

Trend 5: Standardized Prompt Frameworks

 What: Industry-standard ways to structure prompts
 Current: Each organization develops own approaches
 Future: Universal standards emerge (like HTML for web)
 Impact: Easier sharing and reuse of prompts across teams and organizations
 Example: Healthcare industry adopts standardized prompt format for diagnostic
assistance

Trend 6: Context Caching and Memory

 What: AI maintaining context and memory of previous interactions
 Current: Each prompt is independent
 Future: Continuous conversation where AI remembers everything
 Impact: More natural, efficient communication
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Example: "Remember my company details from earlier? Now apply that to..."

Trend 7: Integration with Traditional Programming

 What: Prompts embedded directly in code
 Current: Separate from software development
 Future: Seamless integration where prompts are first-class citizens in programming
 Impact: AI becomes natural part of application development
 Example: Python code like [Link](prompt=my_prompt)

Predictions for 5-10 Years:

 Prompt engineering will become primary skill for many roles
 AI systems will be more intuitive about prompt understanding
 Specialized prompt "operating systems" will emerge
 Cross-platform prompt standards will develop
 Jobs specifically for "prompt architects" will proliferate

Unit 2: Prompt Engineering Techniques

for ChatGPT
2.1 Introduction to Prompt Engineering
Techniques
Why Techniques Matter:

Not all prompts are created equal. Techniques are specific methods that have been
scientifically tested and proven to improve AI output quality.

Think of techniques as tools in a toolbox:

 Each tool serves a specific purpose

 Different tasks require different tools
 Skilled craftsperson knows which tool to use when
 Similarly, effective prompt engineers know which technique to apply

2.2 Instructions Prompt Technique

Core Principle:
Give the AI explicit, clear instructions about what to do, how to do it, and what to avoid.

Basic Structure:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

ROLE: [Who/what should AI act as?]

TASK: [What exactly should it do?]

FORMAT: [How should output be formatted?]

CONSTRAINTS: [What limitations or rules apply?]

AVOID: [What should it NOT do?]

Example 1 - Simple:

Without Instructions:
Summarize this article about climate change.
[article text]
Result: Generic 3-paragraph summary, possibly too technical

With Instructions:
Role: You are a science educator writing for high school students.

Task: Summarize the following article about climate change.

Format:

 One paragraph only

 Simple language (explain technical terms)
 Include one specific example
Constraints:

 150 words maximum

 Assume reader knows nothing about climate science
 Use analogies to explain concepts
Avoid:

 Complex scientific jargon

 Political opinions
 Predictions about future
[article text]
Result: Perfect for high school audience, accessible, concise

Example 2 - Complex (Email Writing):

ROLE: You are a professional business communications specialist.

TASK: Write a professional email to a client who missed an important deadline for project
deliverables.

FORMAT:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Structure: Greeting, acknowledge issue, explain impact, request action, offer support,
closing
 Tone: Professional but empathetic, not accusatory
 Length: 150-200 words
CONSTRAINTS:

 Must maintain client relationship (they're important customer)

 Be clear about deadline for revised submission
 Offer specific ways to help them meet new deadline
 Include contact information for follow-up
AVOID:

 Blame or criticism
 Overly casual language
 Being too aggressive
 Making assumptions about reasons
OUTPUT FORMAT:
[Show me the email with clear sections labeled]

Key Elements of Good Instructions:

Element Why Important Example

Role Sets perspective and expertise "You are an experienced data
level scientist"
Task Makes clear what's needed "Analyze this dataset and identify
trends"
Format Ensures output fits your needs "Provide output as JSON with these
fields"
Constraints Sets boundaries and "Use only information from these
requirements sources"
Examples Show desired output quality "Here's an example of good output..."
Tone Guides communication style "Use formal, academic tone"
Avoid Prevents undesired behaviors "Don't include marketing language"

2.3 Zero-Shot, One-Shot, and Few-Shot Prompting

These techniques vary by the number of examples provided in the prompt.
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Zero-Shot Prompting
Definition:
Asking AI to perform a task without providing any examples.

Structure:
[Task description]
[No examples provided]

Example:
Classify the following tweet as positive, negative, or neutral:
"This new phone is terrible! Battery dies in 2 hours."

When to Use:

 Simple, well-known tasks

 General knowledge questions
 Tasks AI understands intuitively
Advantages:

 Quick and efficient

 No need to create examples
 Shorter prompt
Disadvantages:

 May produce inconsistent results

 AI might not understand exact requirements
 Lower quality for specialized tasks
Result: Likely correct, but format might be uncertain

One-Shot Prompting
Definition:
Providing one example of input and desired output.

Structure:
[Task description]

Example:
[Input example 1]
[Output example 1]

Now, process:
[New input to process]

Example:
Classify tweets as positive, negative, or neutral. Output only the classification.
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Example:
Tweet: "This restaurant has amazing food!"
Classification: Positive

Now classify this:

Tweet: "This new phone is terrible! Battery dies in 2 hours."

When to Use:

 Tasks with specific format requirements

 When you want to show output style
 Moderately complex tasks
Advantages:

 Shows exact format expected

 Better than zero-shot
 Still relatively concise
Disadvantages:

 One example might not be enough

 May not cover edge cases
 Still some inconsistency possible
Result: More consistent format, but might miss nuances

Few-Shot Prompting
Definition:
Providing multiple examples (typically 3-5) of input and desired output.

Structure:
[Task description]

Examples:
[Input example 1] → [Output example 1]
[Input example 2] → [Output example 2]
[Input example 3] → [Output example 3]

Now process:
[New input to process]

Example:
Classify tweets as Positive, Negative, or Neutral. Output only one word.

Examples:
Tweet: "This restaurant has amazing food!"
Classification: Positive
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Tweet: "The weather is okay, nothing special."

Classification: Neutral

Tweet: "This service is absolutely horrible!"

Classification: Negative

Tweet: "The battery on this phone dies too quickly."

Classification: ?

When to Use:

 Complex classification tasks

 Specialized domains
 When consistency is critical
 New or uncommon tasks
Advantages:

 Highest consistency and accuracy

 Covers edge cases better
 AI "learns" pattern from examples
 Most reliable method
Disadvantages:

 Longer prompts
 Requires creating good examples
 More setup work
Result: Highly consistent, high quality, exact format

Practical Comparison:

Aspect Zero-Shot One-Shot Few-Shot

Prompt length Short Medium Long
Consistency Low Medium High ✓
Accuracy Medium Medium-High High ✓
Setup time Minimal Low Medium
Best for Simple Standard tasks Complex tasks
tasks
Example "Define AI" "Translate to "Classify medical
queries Spanish" symptoms"
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Expert Tip - Few-Shot Best Practices:

1. Diverse examples – Show different types of inputs

2. Correct answers – Examples must be accurate
3. Relevant examples – Match the domain/task
4. Proper formatting – Exactly how you want outputs
5. Edge cases – Include unusual/tricky examples
6. Balance – 3-5 examples is usually optimal

2.4 Self-Consistency Prompt

Core Idea:
Ask AI to solve the problem multiple times independently, then combine all answers
to get the best response.

Why It Works:

 Different solution paths can lead to same answer

 If multiple independent attempts agree → likely correct
 Reduces errors and increases confidence
 Especially powerful for reasoning tasks
Basic Structure:

[Problem statement]

Let's think about this step by step. Please solve this problem three times independently, using
different approaches. Then tell me which answer appears most consistent.

Attempt 1:
[Your first approach]

Attempt 2:
[Your second approach]

Attempt 3:
[Your third approach]

Final Answer:
[The most consistent/correct answer]

Example - Math Problem:

Problem: A train leaves Station A at 10 AM traveling at 60 mph. Another train leaves Station
B at 11 AM traveling at 80 mph. The stations are 200 miles apart. When do the trains meet?

Let's solve this three different ways:

Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Approach 1 [Using time equations]

[Work through problem]
Answer: 12:15 PM

Approach 2 [Using relative speed]

[Work through problem]
Answer: 12:15 PM

Approach 3 [Using distance calculations]

[Work through problem]
Answer: 12:15 PM

Since all three approaches give 12:15 PM, that's the correct answer.

When to Use:

 Complex reasoning tasks

 Math and logic problems
 Decision-making scenarios
 When accuracy is critical
Advantages:

 More reliable answers

 Shows reasoning process
 Confidence-boosting
 Catches errors naturally
Disadvantages:

 Much longer response

 Takes more processing time
 Costs more (if using paid API)
 Overkill for simple tasks
Real-World Application:

Company situation: Deciding whether to launch product in new market

Let's analyze this market entry decision three ways:

Business analysis: [Evaluate market size, competition, profitability]

Risk analysis: [Evaluate risks: regulatory, financial, operational]
Strategic analysis: [Evaluate alignment with company goals, competitive advantage]

If all three recommend "launch," confidence is high.

If they conflict, we need more analysis.

2.5 Chain-of-Thought (CoT) Prompting

Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Core Principle:
Ask AI to break complex problems into steps and explain reasoning before giving
final answer.

Why It Works:

 Forces logical thinking

 Makes reasoning transparent
 Catches errors in logic
 Improves accuracy significantly
Basic Structure:

[Problem]

Let's think through this step by step:

Step 1: [First logical step]

Step 2: [Second logical step]
Step 3: [Next logical step]
...
Final Answer: [Conclusion]

Example 1 - Logic Problem:

Without Chain-of-Thought:
Problem: Mary is 6 years older than John. In 2 years, Mary will be twice as old as John. How
old are they now?
Answer: ???

With Chain-of-Thought:
Problem: Mary is 6 years older than John. In 2 years, Mary will be twice as old as John. How
old are they now?

Let's think step by step:

Step 1: Define variables

 Let J = John's current age

 Let M = Mary's current age
 We know: M = J + 6
Step 2: Use the second condition

 In 2 years: John will be J + 2, Mary will be M + 2

 Given: M + 2 = 2(J + 2)
Step 3: Substitute M = J + 6 into the equation

 (J + 6) + 2 = 2(J + 2)
 J + 8 = 2J + 4
Step 4: Solve for J
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 8 - 4 = 2J - J
 4=J
 John is 4 years old
Step 5: Find Mary's age

 M = J + 6 = 4 + 6 = 10
 Mary is 10 years old
Verification:

 Mary is 6 years older than John: 10 = 4 + 6 ✓

 In 2 years: Mary is 12, John is 6. Is 12 = 2 × 6? ✓
Final Answer: John is 4 years old, Mary is 10 years old.

Example 2 - Business Decision:

Problem: Should our company invest in AI training for all employees?

Let's analyze step by step:

Step 1: Understand the investment

 Cost: $50,000 for training program

 Timeline: 3 months
 Participants: 100 employees
Step 2: Consider benefits

 Increased productivity: employees work more efficiently with AI tools

 Competitive advantage: better positioned than competitors
 Employee satisfaction: feels invested in their growth
 Innovation: employees might find new uses for AI
Step 3: Assess risks

 Not all employees may adopt AI effectively

 ROI timeline unclear
 Requires ongoing support and updates
Step 4: Calculate potential ROI

 If 70% adoption rate (70 of 100 employees)

 If each gains 5% productivity = 3.5% company-wide improvement
 With $500,000 annual revenue per employee = $175,000 total benefit
 Payback period: $50,000 / $175,000 = ~3.4 months
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Step 5: Compare to alternatives

 Not training anyone: no investment, but fall behind competitors

 Training select team only: less expensive but benefits fewer people
 Buy AI software without training: software underutilized
Step 6: Conclusion
Investment provides positive ROI in < 4 months, plus long-term competitive benefits.

Final Recommendation: YES, invest in training.

When to Use CoT:

 Complex reasoning problems

 Multi-step decisions
 Explain how AI reached conclusion
 Verify logical correctness
 Educational contexts
Advantages:

 Significantly improves accuracy

 Shows reasoning (explainability)
 Catches logical errors
 Easy to verify conclusions
Disadvantages:

 Longer response
 Takes more processing
 Some tasks don't need it
 Can seem verbose

2.6 Additional Advanced Techniques

Technique: "Explain Before Answering"
Structure:
[Complex problem]

Explain why you might be uncertain about this, then provide your best answer.

Effect:
Shows AI's reasoning and confidence level.
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Technique: "Prompt Chaining"

Concept:
Use output of one prompt as input to another prompt.

Example:
Prompt 1: Analyze the pros and cons of remote work
Output: [List of pros and cons]

Prompt 2 (using Prompt 1's output):

Given these pros and cons, write a policy recommendation for a tech company.

Benefit:
Breaking complex tasks into smaller steps often produces better results.

Unit 3: Prompts for Creative Thinking

3.1 Introduction to Creative Prompts
What is Creative Thinking?

Creative thinking is the ability to generate novel ideas, see unique connections, and
imagine new possibilities not immediately obvious.

Role of Prompts in Creativity:

Good creative prompts:

 Break conventional thinking patterns

 Encourage exploration of multiple perspectives
 Combine ideas in unexpected ways
 Challenge assumptions
 Spark imagination and innovation
Key Principle:

"Prompts don't create creativity—they unleash it."

3.2 Unlocking Imagination and Innovation

Technique 1: Perspective-Shifting Prompts

Concept:
Ask AI to view problem from completely different perspectives.

Example Prompts:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Original problem: How can we improve our app's user engagement?

Shift 1 - Customer perspective:

"If you were a frustrated customer who rarely uses our app, why? What would make you use it
more?"

Shift 2 - Competitor perspective:

"If you were designing our competitor's app, what features would you add that we're
missing?"

Shift 3 - Future perspective:

"Imagine it's 5 years in the future. Our engagement is 10x better. What changed?"

Shift 4 - Beginner perspective:

"You've never used a mobile app before. What confuses you about this app's design?"

Benefits:

 Reveals blind spots

 Generates truly novel ideas
 Combines different viewpoints
 Leads to breakthrough solutions

Technique 2: Constraint-Based Creativity

Concept:
Limit resources/options to spark creative solutions.

Examples:

Scenario 1: Marketing with zero budget

"How would you market our product if we had $0 marketing budget but unlimited
creativity?"

Scenario 2: Designing with extreme limitations

"Design a sustainable product using only recycled materials, with production cost under $5"

Scenario 3: Problem-solving with constraints

"Improve customer satisfaction without increasing staff or costs"

Why It Works:

 Constraints force out-of-box thinking

 Unlimited options lead to paralysis
 Limitations spark creative solutions
 Tests feasibility of ideas

Technique 3: Analogical Thinking

Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Concept:
Solve problems by finding similar patterns in different domains.

Example Prompt:

Problem: How can a hospital reduce patient wait times?

Let's think of analogies:

 How does McDonald's serve hundreds of people quickly?

 How does an airport handle peak passenger flow?
 How does an emergency room prioritize life-threatening cases?
 How does Amazon process millions of orders?
Insights from McDonald's: Assembly line efficiency, staff multi-tasking, prepared
components
Insights from airport: Parallel processing, clear signage, queue management
Insights from emergency: Triage system, priority-based sorting

How could hospital apply these principles?

Power:

 Borrows proven solutions from other domains

 Generates fresh ideas
 Combines best practices

Technique 4: Reverse Brainstorming

Concept:
Instead of "How to improve X," ask "How to make X worse?" Then reverse the ideas.

Example:

How to make customer experience WORSE:

1. Make checkout process complicated

2. Use unclear product descriptions
3. Hide customer service contact
4. Don't respond to customer inquiries
5. Make returns difficult
Reverse these to improve:

1. Streamline checkout to 2 clicks

2. Write crystal-clear product descriptions
3. Prominently display customer service
4. Respond to inquiries within 1 hour
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

5. Make returns free and easy

Result:
Concrete improvement ideas.

Technique 5: Morphological Analysis

Concept:
Break problem into dimensions, explore each dimension's variations.

Example - Designing a new educational product:

Dimension 1: Learning format

 Options: Video, interactive, gamified, hands-on, text-based

Dimension 2: Pace

 Options: Self-paced, group-paced, instructor-led, deadline-driven

Dimension 3: Duration

 Options: 1 hour, 1 day, 1 week, 1 month, ongoing

Dimension 4: Cost

 Options: Free, pay-per-course, subscription, enterprise license

Dimension 5: Platform

 Options: Mobile, web, desktop, offline, VR

Unique combination: Gamified + self-paced + 1-week modules + subscription + mobile app
Result: Engaging, flexible learning app for busy professionals

Power:

 Systematically explores combinations

 Finds unique opportunities
 Comprehensive analysis

Unit 4: Prompts for Effective Writing

4.1 Introduction to Effective Writing Prompts
What Makes Writing Effective?

Effective writing:

 Communicates clearly
 Engages the reader
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Achieves its purpose

 Uses appropriate tone and style
 Removes ambiguity
Role of Prompts in Writing:

Good writing prompts:

 Define audience and purpose

 Set tone and style
 Structure content logically
 Provide examples
 Ensure consistency

4.2 Igniting the Writing Process with Prompts

Phase 1: Pre-Writing Prompts (Planning)

Purpose: Organize thoughts before writing

Prompts:

1. Clarifying the topic:

"Who is my audience?"
"What is the main message I want to convey?"
"What does the reader need to know?"

2. Generating ideas:
"What are 10 aspects of this topic?"
"What questions might my reader have?"
"What are common misconceptions?"

3. Structuring content:
"What's the most logical order?"
"What should I emphasize?"
"What examples would help?"

Phase 2: Drafting Prompts (Creating)

Purpose: Generate initial content

Example Prompt for Blog Post:

You are an experienced tech blogger writing for professionals.

Topic: Machine Learning for Business Leaders

Requirements:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Length: 1,500 words

 Audience: Senior executives with limited ML knowledge
 Tone: Informative but accessible
 Structure:
a. Hook (explain why executives should care)
b. What is ML? (simple explanation)
c. Business applications (3-4 concrete examples)
d. Implementation considerations (ROI, timeline, skills needed)
e. Next steps (what to do now)
Include:

 Real-world company examples

 Practical insights
 Clear explanations of technical concepts

Phase 3: Revision Prompts (Improving)

Purpose: Enhance and refine content

Prompts:

1. Clarity revision:
"Simplify this paragraph so a 10th grader could understand it"
"Is this sentence clear or confusing? How would you rewrite it?"

2. Tone adjustment:
"Make this more professional/casual/humorous"
"This sounds too formal. How would you say it conversationally?"

3. Structure improvement:
"This section feels disorganized. How would you reorganize it?"
"What should come first? Reorganize this logically"

4. Engagement boost:
"Add a compelling example to illustrate this point"
"Make this section more engaging for readers"

Example Writing Process with Prompts:

Step 1 - Planning:
Prompt: I'm writing a guide on "How to Learn AI". Help me outline the key sections.

Step 2 - Drafting:
Prompt: You are an AI education expert. Write the introduction to an AI learning guide.
Make it engaging, explain why learning AI matters, and preview what readers will learn.
Requirements: 300 words, accessible language, motivating tone.
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

Step 3 - Revision:
Prompt: This introduction is too technical. Rewrite it for someone with no programming
experience.
Use simple analogies to explain concepts. Keep it motivating.

Step 4 - Polishing:
Prompt: How would you improve this introduction for clarity and engagement?
What's missing? What should be emphasized more?

Specific Writing Prompts by Type:

For Technical Writing:

You are a technical writer. Explain [complex concept] for [target audience].

Requirements:

 Avoid jargon or define it clearly

 Include a real-world analogy
 Add step-by-step explanation if relevant
 Include a practical example
 Use simple language (8th grade reading level)
For Marketing Copy:
You are a copywriter. Write [marketing material] for [product/service].

Requirements:

 Target audience: [demographics]

 Key benefits: [list]
 Call to action: [desired action]
 Tone: [professional/casual/urgent]
 Avoid: [list what NOT to include]
For Academic Writing:
You are an academic writer. Write [assignment] on [topic].

Requirements:

 Academic level: [e.g., undergraduate, graduate]

 Format: [e.g., essay, research paper]
 Length: [word count]
 Citation style: [e.g., APA, MLA]
 Use credible sources
 Include thesis statement
 Support claims with evidence
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

For Creative Writing:

You are a [type of writer]. Write [type of content] about [topic].

Requirements:

 Genre: [e.g., sci-fi, mystery, romance]

 Tone: [e.g., humorous, dark, inspirational]
 Length: [word count]
 Characters: [describe]
 Setting: [describe]
 Theme: [central idea]

4.3 Common Writing Scenarios with Prompts

Scenario 1: Professional Email

You are a professional communicator. Write an email requesting a meeting.

Details:

 Recipient: Potential business partner

 Purpose: Discuss partnership opportunity
 Tone: Professional, friendly
 Length: 150-200 words
Requirements:

 Clear subject line

 Compelling opening
 Specific meeting times (3 options)
 Clear call to action

Scenario 2: Social Media Post

You are a social media manager for [company/brand].

Write a LinkedIn post about [topic].

Requirements:

 Length: 250 words maximum

 Tone: Professional but personable
 Engagement tactics: Ask question at end
 Include relevant hashtags (3-5)
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Call to action: What should readers do?

 Visual description: What image would work?

Scenario 3: Product Description

You are an e-commerce copywriter. Write a product description.

Product: [Name and type]

Target audience: [Who buys this?]
Key features: [List main features]
Unique benefits: [What makes it special?]

Requirements:

 100-150 words
 Highlight 3-4 key benefits
 Include at least one customer pain point it solves
 Include a call to action (e.g., "Buy now")
 Make it compelling and scannable

Scenario 4: Job Description

You are an HR professional. Write a job description for [position].

Company: [Company type/industry]

Department: [Department]
Key responsibilities: [List main duties]
Required qualifications: [List skills]
Nice-to-have skills: [Bonus skills]

Requirements:

 Clear, engaging tone

 Explain why this role matters
 Include compensation range
 List benefits
 Include application instructions

Exam Preparation Guide

2-Mark Questions
Likely Questions:
Introduction to AI and Applications (1BAIA103/203) Keertipriya Bellary

 Define prompt engineering

 What is zero-shot prompting?
 Name two prompt types
 What is Chain-of-Thought?
 Explain role-based prompts
 What is few-shot prompting?
Study Format: 1-2 sentence definition + one-line explanation

5-Mark Questions
Likely Questions:

 Explain prompt engineering with an example

 Compare zero-shot, one-shot, and few-shot prompting
 Describe self-consistency prompting with example
 How does chain-of-thought improve AI responses?
 What are advantages of prompt engineering?
Study Format: Definition → Key points (2-3) → Example

10-Mark Questions
Likely Questions:

 Comprehensive overview of prompt engineering evolution and techniques

 Compare different prompt techniques and when to use each
 How do creative thinking prompts work?
 Explain effective writing prompt strategies
 Discuss future of LLM communication
Study Format: Introduction → Detailed explanation with examples → Comparative
analysis → Conclusion