Basic Prompting Techniques

Now that you understand the fundamentals of prompt engineering, let's dive into the core prompting techniques that will form the foundation of your prompt engineering skills.

What You'll Learn

By the end of this lesson, you'll be able to:

• ✅ Master zero-shot, one-shot, and few-shot prompting techniques
• ✅ Apply chain-of-thought prompting for complex reasoning
• ✅ Use multimodal prompting with images and text
• ✅ Implement function calling and RAG techniques
• ✅ Choose the right technique for different scenarios

Zero-Shot Prompting

Zero-shot prompting is when you ask an AI to perform a task without providing any examples. The model relies entirely on its pre-trained knowledge.

When to Use Zero-Shot

• Simple, straightforward tasks
• Common knowledge questions
• Basic text generation
• When you want quick results

Examples

One-Shot Prompting

One-shot prompting provides a single example to guide the AI's response. This helps the model understand the expected format and style.

When to Use One-Shot

• When you need specific formatting
• For tasks that benefit from examples
• When zero-shot isn't giving desired results
• To establish a particular style or tone

Examples

Few-Shot Prompting

Few-shot prompting provides multiple examples to establish a clear pattern. This is especially effective for complex tasks or when you need consistent formatting.

When to Use Few-Shot

• Complex tasks requiring multiple examples
• When you need consistent output format
• For tasks with multiple possible approaches
• When one-shot isn't sufficient

Examples

Chain-of-Thought Prompting

Chain-of-thought (CoT) prompting encourages the AI to show its reasoning process step by step. This is particularly useful for complex problem-solving tasks.

When to Use Chain-of-Thought

• Complex reasoning problems
• Mathematical calculations
• Logical puzzles
• When you want to understand the AI's thinking process
• Debugging and troubleshooting

Examples

Advanced Prompting Techniques

Multimodal Prompting

Multimodal prompting involves using multiple types of input (text, images, audio, video) to get more comprehensive responses from AI models.

When to Use Multimodal

• When you need to analyze visual content
• For tasks requiring both text and image understanding
• When working with documents, charts, or diagrams
• For creative tasks involving multiple media types

Examples

Function Calling

Function calling allows AI models to execute specific functions or API calls based on your prompts, enabling more interactive and dynamic responses. The Model Context Protocol (MCP) provides a standardized approach to function calling across different AI platforms.

When to Use Function Calling

• When you need real-time data
• For tasks requiring external API integration
• When you want the AI to perform actions
• For dynamic content generation

Examples

RAG (Retrieval-Augmented Generation)

RAG combines AI generation with external knowledge retrieval, allowing models to access up-to-date information beyond their training data.

When to Use RAG

• When you need current information
• For domain-specific knowledge
• When working with proprietary data
• For tasks requiring recent events or data

Examples

Prompt Templates

Here are some useful templates you can adapt for different tasks:

💡 Pro Tip: For a comprehensive library of ready-to-use templates, check out our Prompt Templates Library lesson.

Analysis Template

Creative Writing Template

Problem-Solving Template

Best Practices for Each Technique

Zero-Shot Best Practices

• Be very specific about what you want
• Use clear, unambiguous language
• Consider the model's knowledge cutoff
• Test with different phrasings

One-Shot Best Practices

• Choose a representative example
• Ensure the example matches your desired output format
• Keep examples concise but complete
• Use examples that demonstrate the right level of detail

Few-Shot Best Practices

• Provide 2-4 examples for best results
• Ensure examples are consistent in format and style
• Vary the examples to show different scenarios
• Make sure examples are relevant to your task

Chain-of-Thought Best Practices

• Explicitly ask for step-by-step reasoning
• Use phrases like "Let's solve this step by step" or "Think through this"
• Provide enough context for the reasoning process
• Be patient with longer responses

Practice Exercises

Exercise 1: Zero-Shot

Exercise 2: One-Shot

Exercise 3: Few-Shot

Exercise 4: Chain-of-Thought

Common Pitfalls to Avoid

1. Inconsistent Examples: Make sure your few-shot examples follow the same pattern
2. Overly Complex Prompts: Start simple and add complexity as needed
3. Ignoring Context: Provide enough context for the AI to understand the task
4. Unrealistic Expectations: Don't expect perfect results from basic techniques alone
5. Poor Formatting: Use clear formatting to separate examples and instructions

Summary

In this lesson, you've mastered the fundamental prompting techniques:

• Zero-Shot Prompting: Direct instructions without examples
• One-Shot Prompting: Single example to establish format and style
• Few-Shot Prompting: Multiple examples to establish patterns
• Chain-of-Thought: Step-by-step reasoning for complex problems
• Multimodal Prompting: Combining text with images and other media
• Function Calling: Enabling AI to execute specific functions
• RAG: Retrieval-augmented generation for current information

Self-Check

Test your understanding of basic prompting techniques:

1. Which technique is best for simple, straightforward tasks?

   • Zero-shot prompting
   • One-shot prompting
   • Few-shot prompting

2. When should you use chain-of-thought prompting?

   • For simple questions
   • For complex reasoning problems
   • For creative writing tasks

3. What's the main advantage of few-shot prompting?

   • It's faster than other techniques
   • It establishes clear patterns and consistency
   • It requires less thinking

4. Which technique combines AI generation with external knowledge?

   • Function calling
   • RAG (Retrieval-Augmented Generation)
   • Multimodal prompting

5. What should you avoid when using one-shot prompting?

   • Providing clear examples
   • Using examples that don't match your desired output
   • Being specific about requirements

Next Steps

In the next lesson, you'll learn about LLM Configuration and how to optimize your prompts for different AI models like GPT-5, Claude 4, and Gemini 2.5.

📚 Quick Reference: Need a fast reminder of these techniques? Check out our Prompt Engineering Cheat Sheet for a quick reference guide.

Practice Time! Try these techniques with different AI models and see how they respond. Remember, practice makes perfect in prompt engineering!