Welcome! If you think “prompt” is just the name of a detergent or that “generative AI” is a new diet, you’re in the right place. Here I’ll explain, without jargon and without panic, how to start talking to generative artificial intelligences (like ChatGPT, Copilot, Gemini and company) and get responses that don’t sound like they were written by a refrigerator.

What is generative AI? (And why should you care?)

Imagine a robot that, instead of bringing you coffee, answers your questions, writes texts, invents stories, suggests recipes and maybe even helps with homework. That’s generative AI: software that creates new content based on your requests (the famous “prompts”).

You don’t need to be an engineer or computer wizard: you just need to know how to write a question. The trick? Knowing how to write it well!

What is a prompt? (Spoiler: it’s not black magic)

A prompt is simply your request to the AI. It can be a question, a command, a sentence. The clearer the prompt, the better the AI understands what you want. Think of it like ordering at a restaurant: if you just say “food,” the waiter gives you a weird look. If you say “a margherita pizza without anchovies, please,” you’re already a prompt engineer!

The golden rules of prompt engineering (easy version)

Here are some tricks to avoid getting “I don’t understand the question” responses:

1. Be clear and specific 🧐

• ❌ Bad: “Tell me about dogs.”
• ✅ Good: “Explain to me in simple terms how to train a Labrador puppy.”

2. Give some context 🏞️

• ❌ Bad: “Write a recipe.”
• ✅ Good: “Write an easy chocolate cake recipe for someone who can’t cook.”

3. Ask for the format you want 📋

• “Give me 3 reasons why I should use AI.”
• “Summarize this text in 5 lines.”

4. Don’t be afraid to ask simple questions! 😅

The AI doesn’t judge (it gets confused at most). If you don’t know where to start, ask basic things. Example: “What is generative AI? Explain it to me like I’m 10 years old.”

Practical examples (so you never get it wrong)

• “Write a polite email to ask your boss for time off.”
• “Explain the difference between cloud and fog, but without using technical terms.”
• “Make up a joke about programmers.”

Common mistakes (and how to avoid them)

• Prompts that are too vague: “Tell me something.” (Response: “Something.”)
• Prompts that are too long and confusing: “I’d like to know everything about everything, but also nothing, maybe something interesting.” (The AI has an existential crisis.)
• Forgetting the audience: if you want an explanation for children, say so!

A pro trick: try, fail, try again! 🔄

There is no perfect prompt on the first try. Change the words, add details, ask the AI to rewrite or explain better. It’s like chatting with a slightly distracted friend: sometimes you need to guide them.

Conclusion

Prompt engineering is not an occult science: it’s just a matter of clarity, patience and a bit of irony. The more you practice, the more the AI understands you (and the less you risk getting refrigerator-style responses).

Have fun and… may the prompt be with you! 🚀

---

Want to go deeper?

If this guide seemed too “for dummies” and you want to discover more advanced techniques, professional strategies and all the tricks of the trade, also read the dedicated article:

👉 Prompt Engineering: Complete Guide and Best Practice

So you can go from beginner level to expert without losing your smile! 😎