Chapter 2: Types of Prompts
In Chapter 1, we explored why clear prompts matter. Now let’s go deeper and look at the main prompt styles you can use. Each style has its own personality — picking the right one can completely change how the model responds.
Instruction Prompts
This is the simplest style: you tell the AI what to do in plain language. Perfect when you know exactly what you want.
Example:
Few-Shot Prompts
Few-shot prompts give the model a couple of examples before asking for the real answer. This is like showing a friend how to solve one problem, then letting them try the next one themselves.
Example:
Product: A coffee delivery app
Name: BrewBuddy
Product: A budgeting tool for teens
Name: MoneyMaven
Product: A language learning game
Name:
Chain-of-Thought Prompts
These encourage the model to “show its work.” Instead of just giving the answer, the AI explains its reasoning step by step — useful for problem-solving or analysis.
Example:
Role-Based Prompts
You can ask the AI to “pretend” to be someone — a teacher, a chef, a marketing expert — to get more focused answers. This helps the AI adopt the right tone and style.
Example:
💡 Tip
🧩 Quick Quiz
1. Which prompt type is best for step-by-step reasoning?
Show Answer
Chain-of-Thought Prompt.
2. Multiple Choice: Which prompt type is most useful if you want the AI to take on a specific tone or persona?
- A. Few-Shot
- B. Role-Based
- C. Instruction
- D. Chain-of-Thought
Show Answer
✅ B. Role-Based — it lets you set a context for how the AI should respond.
3. Reflection: Rewrite a recent question you asked an AI as a role-based prompt. What difference do you expect in the response?