Teaching-materialsPrior-teaching2025-26Ai-workshops-for-teachersNon-technical-ai-workshops-for-teachers4 Comprehensive Prompting Guide

4 Comprehensive Prompting Guide

Effective Strategies for Working with Large Language Models

Introduction

Large Language Models (LLMs) are not search engines. They generate responses through pattern matching and statistical prediction rather than retrieving existing information. Understanding this fundamental difference is crucial for effective prompting.

Key principle: Getting good responses often means adding explanatory context and preference information to your instruction or query. One-line prompts rarely produce optimal results because they lack the information the model needs to understand your specific needs.

The Four-Component Framework

Effective prompts typically include four key components:

1. Goal: What You Want Accomplished

The goal states what you would like the tool to do, create, find, accomplish, address, or answer. Be specific about the desired outcome.

Example goals:

  • "Generate a lesson plan for teaching fractions to adult learners"

  • "Create a professional summary for a LinkedIn profile"

  • "Draft a cover letter connecting my experience to this role"

2. Format: How the Response Should Be Structured

Format specifies the shape, structure, or form the response should take. This includes length, organization, and structural elements.

Format considerations:

  • Should it be bullet points, paragraphs, or a numbered list?

  • How long should it be?

  • Should sources be included?

  • What sections or components should be included?

3. Tone: The Voice and Emotional Quality

Tone describes how the response should sound or feel.

Example tone specifications:

  • "Professional but approachable, avoiding corporate jargon"

  • "Friendly and encouraging, suitable for adult learners"

  • "Academic but accessible, explaining technical concepts clearly"

4. Context: Information Needed to Complete the Task

Context provides the necessary information, background, and circumstances that enable effective completion of the task.

What to include in context:

  • Who is the audience?

  • What is the purpose?

  • Where will this be used?

  • What background information is needed?

  • What constraints exist?

Critical safety note: Never include personal information like addresses, phone numbers, bank details, or other sensitive data in prompts.

Practical Examples

Example 1: Cover Letter

Weak prompt: "Write me a cover letter."

Strong prompt:

Goal: I need a draft cover letter for a customer service position. Context: I have 3 years of retail experience and 2 years in hospitality. The job posting emphasizes conflict resolution and team collaboration. The company is a mid-sized tech startup with a casual but professional culture. Format: Three paragraphs. Approximately 300-350 words total. Tone: Professional but friendly and approachable, not overly formal.

Example 2: Educational Content

Weak prompt: "Explain fractions."

Strong prompt:

Goal: Create a one-page explanation of how to add fractions with different denominators. Context: My students are adult learners (ages 25-45) returning to education. Many have math anxiety. Use everyday examples like measuring ingredients or splitting bills. Format: Start with a simple explanation, then provide 3 worked examples of increasing complexity, then 3 practice problems. Tone: Encouraging and patient. Make it feel achievable, not intimidating.

Official Prompting Guides

Critical Evaluation and Verification

Even with perfect prompts, LLM outputs require critical evaluation. Remember:

  • LLMs can generate false information with complete confidence

  • Citations may be fabricated—verify every source

  • Factual accuracy requires your verification, not the model's

  • You remain responsible for what you submit or publish

Built with LogoFlowershow