06. Use-Case Prompts: Business, Education & Support

67. Image Generation Prompts

🇬🇧 English Definition

Image generation prompts describe subject, composition, environment, lighting, artistic style, camera lens/view, aspect ratio, and explicit negative exclusions.

🇮🇳 Hinglish Definition

Image Generation Prompts: Subject, composition, camera view, lighting, style, and negative prompt (e.g. 'no blur, no extra limbs') specify karein.

// Diffusion Image Generation Prompt
image_prompt = """SUBJECT: A futuristic cyberpunk data center with glowing blue fiber-optic cables.
COMPOSITION: Wide-angle lens shot, symmetrical perspective.
LIGHTING: Neon blue and vibrant orange volumetric lighting.
STYLE: Photorealistic 8K render, Octane Render style.
NEGATIVE PROMPT: blurry, low resolution, crowded, messy wiring."""

print(image_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

SUBJECT & NEGATIVE PROMPT: Specifies target visual elements while explicitly filtering low-quality artifacts.

🎯 Interview Focus & Questions

Q: How do negative prompts improve image generation quality?
A: Negative prompts push the diffusion denoising process away from undesirable latent features (e.g. blurriness, extra fingers, distorted text).

68. Video Generation Prompts

🇬🇧 English Definition

Video generation prompts define subject, temporal action, camera movement (panning, zoom), scene lighting, duration, and motion continuity constraints.

🇮🇳 Hinglish Definition

Video Generation Prompts: Subject, camera motion (pan left, zoom in), action continuity, and lighting specify karein for smooth video frames.

// Video Generation Prompt Structure
video_prompt = """SUBJECT: A futuristic autonomous drone taking off.
ACTION: Drone slowly elevates vertically and pans left across a modern skyline.
CAMERA: Smooth cinematic tracking shot, 24fps.
LIGHTING: Golden hour sunset lighting with lens flare."""

print(video_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

ACTION & CAMERA: Explicitly dictates temporal movement and camera trajectory for video diffusion models.

🎯 Interview Focus & Questions

Q: What is the main challenge when writing prompts for AI video models?
A: Instructing clear, physically plausible camera and object motion without causing temporal warping or frame jitter.

69. Prompting for Business

🇬🇧 English Definition

Business prompts state strategic objectives, target metrics (KPIs), ROI expectations, brand voice, and actionable recommendations tied to measurable outcomes.

🇮🇳 Hinglish Definition

Business Prompts: Objective, target KPI, business constraints, brand voice aur actionable recommendations define karein tied to measurable ROI.

// Business Strategy Prompt
business_prompt = """ROLE: Chief Strategy Officer
TASK: Recommend 3 cost-reduction strategies for a SaaS company.
CONTEXT: Current ARR is $10M; target is reducing cloud infrastructure spending by 20%.
OUTPUT: Executive Memorandum with Strategy Name, Estimated Savings ($), and Implementation Risk."""

print(business_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

OUTPUT: Executive memorandum format anchoring recommendations directly to dollar savings and risk.

🎯 Interview Focus & Questions

Q: How do you prompt an LLM to produce business recommendations suitable for C-suite executives?
A: Request executive summary format, demand quantitative evidence/ROI metrics, and enforce concise, actionable recommendations.

70. Prompting for Education

🇬🇧 English Definition

Education prompts define learner knowledge level, pedagogical progression (concept -> example -> practice -> feedback), misconceptions, and interactive assessment rules.

🇮🇳 Hinglish Definition

Education Prompts: Learner level, pedagogical flow (Concept -> Example -> Practice -> Feedback), and misconception handling specify karein.

// Pedagogical Prompt Flow
edu_prompt = """ROLE: Lead Computer Science Educator
TARGET LEARNER: First-year undergraduate student.
TOPIC: Explain Recursion in Python.

PEDAGOGICAL STEPS:
1. CONCEPT: Explain base case vs recursive step using a simple analogy.
2. CODE EXAMPLE: Provide factorial function with comments.
3. EXERCISE: Ask the student a 1-question practice quiz."""

print(edu_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

PEDAGOGICAL STEPS: Guides student through concept, code example, and interactive practice.

🎯 Interview Focus & Questions

Q: How do you instruct an AI tutor prompt to encourage active learning?
A: Instruct the model: 'Do not provide direct code solutions immediately. Ask guiding Socratic questions to lead the student to the answer.'

71. Prompting for Interviews

🇬🇧 English Definition

Interview prompts specify target job role, candidate experience level, technical domain, interview style, and evaluation rubrics for grading candidate answers.

🇮🇳 Hinglish Definition

Interview Prompts: Target role (e.g. Senior Backend Dev), interview format (Mock Interview), and answer scoring rubric specify karein.

// Technical Interviewer Prompt
interview_prompt = """ROLE: Principal Systems Architect conducting a technical interview.
POSITION: Senior Python Developer (5+ years exp).

INSTRUCTION: Ask ONE question at a time about distributed system caching.
Wait for candidate answer. Evaluate answer against Rubric (0-5) before asking next question."""

print(interview_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

ONE question at a time: Prevents dumping multiple questions simultaneously in conversational mock interviews.

🎯 Interview Focus & Questions

Q: Why is turn-by-turn state management essential for mock interview prompts?
A: It maintains conversational flow, allowing the model to adapt follow-up questions dynamically based on the candidate's previous response.

72. Prompting for Resume / Job Applications

🇬🇧 English Definition

Resume prompts take candidate factual experience, target job descriptions, and strict anti-hallucination rules prohibiting fabricated metrics or job titles.

🇮🇳 Hinglish Definition

Resume Prompts: Actual experience and target Job Description supply karein. Strict rule: Model real facts ko fabricate or fake na kare.

// Resume Bullet Optimizer Prompt
resume_prompt = """TASK: Rewrite candidate resume bullet points for a Senior DevOps position.
CONSTRAINT: Rely ONLY on the provided work history. Do NOT invent companies, degrees, or fake metrics.

Original: "Managed Docker containers."
Target: XYZ Action Verb format with real metrics.
Rewritten:"""

print(resume_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

CONSTRAINT: Strict anti-fabrication rule protecting candidate integrity.

🎯 Interview Focus & Questions

Q: How do you prevent LLMs from inventing fake employment history on resumes?
A: Add strict constraints: 'Use ONLY facts stated in . If a metric is unstated, do not invent numbers.'

73. Prompting for Marketing / SEO

🇬🇧 English Definition

Marketing prompts specify target persona, platform rules, product value proposition, primary/secondary keywords, tone, and strict anti-keyword-stuffing constraints.

🇮🇳 Hinglish Definition

Marketing/SEO Prompts: Target persona, platform (LinkedIn/Blog), primary keywords, tone, and 'no keyword stuffing' constraint specify karein.

// SEO Blog Outline Prompt
seo_prompt = """ROLE: Content Marketing Lead
TOPIC: Best Practices for LLM Prompt Engineering
PRIMARY KEYWORD: "prompt engineering techniques"
AUDIENCE: Software Developers & Product Managers
OUTPUT: H1, H2, H3 Outline with Meta Description (under 160 chars)."""

print(seo_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

OUTPUT: Requests structured H1-H3 blog outline and character-constrained meta description.

🎯 Interview Focus & Questions

Q: How do you instruct an LLM to generate SEO content without violating search engine quality guidelines?
A: Instruct: 'Focus on high-value, original developer insights. Integrate keywords naturally without keyword stuffing or fluff.'

74. Prompting for Customer Support

🇬🇧 English Definition

Customer support prompts provide verified policy documents, escalation rules, allowed refund limits, brand empathy tone, and strict fallback procedures.

🇮🇳 Hinglish Definition

Customer Support Prompts: Knowledge base context, escalation rules, allowed refund limits, and fallback guidelines specify karein.

// Customer Support Bot Prompt
support_prompt = """SYSTEM POLICY: Customer Support Assistant for E-Commerce.
ALLOWED REFUNDS: Up to $30 for damaged items with photo proof.

Rule: If customer requests refund > $30, respond: "I am transferring your request to a manager."
Rely ONLY on the provided FAQ context below."""

print(support_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

ALLOWED REFUNDS & Rule: Enforces concrete policy boundary ($30 ceiling) and escalation fallback.

🎯 Interview Focus & Questions

Q: Why must customer support prompts explicitly state allowed policy limits?
A: To prevent the AI from making unauthorized promises (e.g. promising 100% refunds or free products) that legally bind the company.

75. Prompting for Document Extraction

🇬🇧 English Definition

Document extraction prompts define fields, page references, confidence threshold rules, and strict schema formats to pull structured data from PDFs/scans.

🇮🇳 Hinglish Definition

Document Extraction Prompts: Target fields, page number references, and strict JSON schemas define karein for PDF data extraction.

// Document Extraction Prompt Template
doc_extract_prompt = """Extract invoice details from the document text.

JSON SCHEMA:
{
    "invoice_id": "string",
    "vendor_name": "string",
    "total_amount": float,
    "line_items": [{"description": "string", "amount": float}]
}

Rule: If a field is unreadable, set to null. Include page_number where found."""

print(doc_extract_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

line_items & page_number: Handles nested array extraction while preserving document traceability.

🎯 Interview Focus & Questions

Q: How do you handle OCR errors in document extraction prompts?
A: Instruct the model: 'Correct obvious OCR character typos (e.g. $1O0 -> $100) based on context, but do not guess missing numbers.'

76. Prompting for Emails

🇬🇧 English Definition

Email prompts specify recipient role, relationship, core purpose, key facts, call-to-action (CTA), tone, and desired length constraints.

🇮🇳 Hinglish Definition

Email Prompts: Recipient, core objective, key facts, CTA (Call To Action), and professional tone specify karein.

// Professional Email Prompt
email_prompt = """TASK: Write a follow-up email to a enterprise client after a product demo.
RECIPIENT: CTO of a Fintech company.
KEY FACTS: Demoed RAG Security module; client requested SOC2 compliance report.
CTA: Schedule 15-min security review call next Tuesday.
TONE: Professional, concise, respectful. Under 150 words."""

print(email_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

KEY FACTS & CTA: Anchors email content to specific demo outcomes and requests clear meeting follow-up.

🎯 Interview Focus & Questions

Q: What constraint ensures generated business emails remain concise?
A: Specify length limits ('Under 150 words') and mandate a single explicit Call to Action (CTA).

77. Prompting for Presentations

🇬🇧 English Definition

Presentation prompts request slide-by-slide outlines, slide titles, key bullet points, visual layout guidance, and speaker notes.

🇮🇳 Hinglish Definition

Presentation Prompts: Slide-by-slide outline (Title, Bullets, Visual Ideas, Speaker Notes) generate karne ki instruction dein.

// Slide Deck Outline Prompt
deck_prompt = """TASK: Create a 5-slide presentation outline on "Introduction to Vector DBs".

FOR EACH SLIDE OUTPUT:
- Slide Title
- 3 Bullet Points (max 10 words each)
- Visual Concept Idea
- Speaker Notes (2 sentences)"""

print(deck_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

FOR EACH SLIDE: Standardizes structure across title, concise bullets, visual guidance, and speaker notes.

🎯 Interview Focus & Questions

Q: How do you prompt an LLM to generate slides that are visually uncluttered?
A: Enforce strict bullet constraints ('Maximum 3 bullets per slide, under 10 words per bullet').

78. Prompting for Learning / Tutoring

🇬🇧 English Definition

Tutoring prompts instruct the model to provide beginner-friendly explanations, real-world analogies, code walkthroughs, practice exercises, and self-assessments.

🇮🇳 Hinglish Definition

Tutoring Prompts: Simple explanation, real-world analogy, step-by-step code walkthrough, and a practice quiz demand karein.

// Socratic AI Tutor Prompt
tutor_prompt = """ROLE: Patient AI Coding Tutor
TOPIC: Explain Python Decorators.

STRUCTURE:
1. Simple Analogy (Explain using a real-world wrapper example)
2. Minimal Code Example
3. Line-by-Line Code Breakdown
4. 1-Question Quiz (Wait for student response before revealing answer)"""

print(tutor_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

STRUCTURE 1-4: Combines analogy, code, line-by-line breakdown, and interactive quiz.

🎯 Interview Focus & Questions

Q: Why is Socratic tutoring effective in AI prompts?
A: Socratic prompts ask guiding questions, forcing the learner to engage actively rather than passively reading solutions.

79. Prompting for Code Review

🇬🇧 English Definition

Code review prompts instruct the model to review code for correctness, security vulnerabilities (OWASP), performance bottlenecks, maintainability, and test coverage.

🇮🇳 Hinglish Definition

Code Review Prompts: Code ko Security (OWASP Top 10), Performance, Readability, aur Edge Cases ke against review karke table mein output dene bolen.

// Automated Code Review Prompt
code_review_prompt = """ROLE: Senior Security & Code Reviewer
TASK: Review the Python function below.

CHECKLIST:
- Security vulnerabilities (SQLi, XSS, Hardcoded secrets)
- Performance bottlenecks (O(N^2) loops)
- Exception handling completeness

OUTPUT: Markdown Table with columns [Issue, Severity, Code Fix]."""

print(code_review_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

CHECKLIST & Markdown Table: Directs code audit toward security and performance vulnerabilities.

🎯 Interview Focus & Questions

Q: How do you instruct a Code Review prompt to minimize false positive warnings?
A: Instruct: 'Flag only High and Critical severity issues. Omit minor subjective style nitpicks.'

80. Prompting for Testing

🇬🇧 English Definition

Testing prompts generate unit tests, boundary test cases, negative test cases, and mock fixtures for specific functions using frameworks like `pytest`.

🇮🇳 Hinglish Definition

Testing Prompts: Function ke liye happy path unit tests, edge-case boundary tests, and negative error tests (`pytest`) generate karein.

// Unit Test Generation Prompt
test_gen_prompt = """TASK: Write pytest unit tests for the Python function below.

FUNCTION:
def divide(a: float, b: float) -> float:
    return a / b

REQUIREMENTS:
1. Test happy path division.
2. Test division by zero exception (pytest.raises(ZeroDivisionError)).
3. Test negative floating-point numbers."""

print(test_gen_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

REQUIREMENTS 1-3: Ensures complete coverage across happy path, edge-case exceptions, and boundary inputs.

🎯 Interview Focus & Questions

Q: What prompt instructions ensure generated unit tests are production-ready?
A: Request explicit framework usage (`pytest`), mandate parameterized test cases (`@pytest.mark.parametrize`), and require negative error assertion tests.

81. Prompting for Documentation

🇬🇧 English Definition

Documentation prompts generate API references, function docstrings (Google/NumPy style), prerequisites, code examples, and error codes.

🇮🇳 Hinglish Definition

Documentation Prompts: Google-style docstrings, API endpoints reference, prerequisites, aur runnable code examples generate karwate hain.

// Google-Style Docstring Generator Prompt
doc_prompt = """TASK: Write Google-style Python docstring for the function below.

FUNCTION:
def calculate_tax(income: float, rate: float = 0.2) -> float:
    return income * rate

OUTPUT FORMAT: Include Args, Returns, Raises, and a runnable Example block."""

print(doc_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

Google-Style Docstring: Enforces standard docstring format with Args, Returns, and Examples.

🎯 Interview Focus & Questions

Q: How do you ensure generated API documentation remains accurate over time?
A: Auto-generate documentation prompts directly from source code ASTs or OpenAPI JSON schemas during CI/CD build pipelines.

82. Prompting for Agents with Memory

🇬🇧 English Definition

Agent memory prompts explicitly segregate long-term persistent user facts from short-term task state, defining what information may be saved vs forgotten.

🇮🇳 Hinglish Definition

Agents with Memory Prompts: Ephemeral session context aur long-term DB persisted facts ke boundary rules define karein.

// Memory Policy System Prompt
memory_agent_prompt = """SYSTEM POLICY: Customer Agent with Memory.

MEMORY RULES:
- REMEMBER: User preferred name, language choice, account ID.
- DO NOT STORE: Credit card details, temporary passwords, one-time OTPs.

Update user memory dict only when explicit new preferences are stated."""

print(memory_agent_prompt)
            

🔍 Line-by-Line Prompt / Code Breakdown

MEMORY RULES: Defines explicit whitelist and blacklist for persistent memory storage.

🎯 Interview Focus & Questions

Q: Why must sensitive data be explicitly blacklisted in agent memory prompts?
A: To prevent private credentials (passwords, PII, API keys) from being permanently stored in long-term vector databases or logs.

83. Prompting for Multi-Agent Systems

🇬🇧 English Definition

Multi-agent prompts give each agent a narrow role, explicit handoff contracts, termination conditions, and shared state schema rules.

🇮🇳 Hinglish Definition

Multi-Agent System Prompts: Har agent ko specific role (e.g. Researcher vs Editor), handoff schema, aur shared state update rules dein.

// Multi-Agent Handoff Protocol Prompt
multi_agent_spec = """AGENT 1: Research Agent (Finds raw data)
HANDOFF CONTRACT: Output raw JSON with keys [query, facts_list].

AGENT 2: Writer Agent (Takes Research Agent JSON)
HANDOFF CONTRACT: Output 300-word Markdown article based ONLY on facts_list.

Do not overlap responsibilities between Agent 1 and Agent 2."""

print(multi_agent_spec)
            

🔍 Line-by-Line Prompt / Code Breakdown

HANDOFF CONTRACT: Defines exact JSON interface for passing control from Agent 1 to Agent 2.

🎯 Interview Focus & Questions

Q: What causes multi-agent systems to break down?
A: Ambiguous responsibility boundaries, unstandardized handoff schemas, and lack of global termination conditions.