Prompts 101

What is a Prompt?

Put simply, a prompt is any text or instruction you give an AI. Examples can be a Simple question: What is the weather in NYC? or a Math problem: What is √4? Or instructions for a specific Task: Write a short ballad about sea creatures 🦑.

Think of this like chatting with a helpful coworker. You explain what you want, and they try to deliver. Prompts can be quick questions or multi-step tasks.

Quick story: I once asked an AI to "plan a dinner." The result was random. When I added my goal (simple), guests (4 people), time (30 minutes), and budget (low), it nailed it. Clear prompts get clearer results.

How to ask well (start simple)

  • Be clear about the goal: "Summarize this article in 3 bullet points."
  • Add helpful context: "I’m new to Python."
  • Set constraints: "Use plain language, no jargon."
  • Optional format: "Return a checklist."
  • One step at a time: "First outline, then write."

Send a prompt. (This can be anything)

🤖 Prompt Tester
System Prompt
Your name is Joe. Joe is an average man with average American tastes.
Model: gpt-4o-mini
0/5 messages used

Try asking the AI what its favorite food is. Then tighten your prompt (goal, context, constraint) and compare the answers.

A quick note on different kinds of prompts

Sometimes apps include behind-the-scenes instructions that guide tone and safety. Those are often called system prompts. We’ll cover the differences in the next lesson.

Common pitfalls (to notice early)

  • Vague in, vague out: “Help me” is too open. Add goal, context, constraints.
  • Too much at once: Split big tasks into steps.
  • Sounds confident, still wrong: Always check facts when it matters. Hallucinations and bias are real concerns in AI; use clear prompts and verify important details. See the guidance on effective prompting and on hallucinations/bias from MIT Sloan’s AI Basics. [1]

Key Takeaways:

  • A prompt is an instruction to the AI.
  • Clarity wins: add goal, context, constraints, and a format when helpful.
  • Iterate: ask, review, refine.
  • System prompts shape behavior behind the scenes-learn them next.

More Resources:

Sources: