Artikel 7
KI Grundlagen

What Is a Prompt? The Art of Good Instructions

Sebastian Rydz10. November 202510 min Lesezeit

What Exactly Is a Prompt?

Imagine you're standing in a restaurant in a foreign country. You're hungry, the menu is in a language you don't speak, and the waiter is patiently waiting. You point at something random and hope for the best. Sometimes you get something delicious. Sometimes you end up with a dish you'd never have ordered.

That's exactly what communicating with AI feels like when you don't know how to put your wishes into words. You type something in, hit Enter, and hope for a useful result. Sometimes it works, sometimes it doesn't. But why is that?

Here's the good news: It's not your fault, and it's not the AI's fault either. It's about the instruction you give. And that instruction has a name: prompt. That's exactly what this article is about. You'll learn what a prompt really is, why it matters so much, and how better wording leads to dramatically better results. Don't worry, we'll start nice and easy.

A prompt is simply your instruction to the AI. Every time you type something into ChatGPT, Claude, or any other AI tool, you're entering a prompt. The word means something like “cue” or “input” - an invitation for the AI to respond. Sounds technical? It really isn't.

Let me paint a picture. Imagine you're giving a task to a very talented but extremely literal assistant. This assistant is incredibly fast, has enormous knowledge, and wants to help you with almost anything. But there's one thing about them: they do exactly what you say. Not more, not less. Not what you might mean. What you actually say.

If you say “Write me something about dogs,” you'll get something about dogs. Maybe a scientific text, maybe a story, maybe a list of dog breeds. The AI has no way of knowing what you had in mind. Your prompt was too open.

If instead you say “Write me a short, friendly paragraph about the benefits of dogs as pets for families with children,” you'll get exactly that. The difference? Your prompt was more precise. And that's what the art of good instructions is all about.

Question or Prompt: The Crucial Difference

Many people use AI tools like a search engine. They type in a question and expect an answer. That works, but it wastes enormous potential. Because AI isn't a search engine. It's a tool that responds to instructions.

Here's the crucial difference between a question and a prompt:

A question: “What can I cook today?”

That's like asking someone on the street: “What should I eat?” The answer will be generic because that person knows nothing about you. What's in your fridge? How many people are you cooking for? Do you have allergies? How much time do you have? All of that remains open.

A prompt: “I have potatoes, onions, and cheese in my fridge. Suggest three quick vegetarian meals I can prepare in under 30 minutes for two people. Please include a short ingredient list and step-by-step instructions.”

See the difference? The prompt includes context (what you have), a clear task (suggest three meals), constraints (vegetarian, under 30 minutes, two people), and a desired format (ingredient list and instructions). The AI now knows exactly what to deliver.

An example from the workplace makes it even clearer:

  • Question: “How do I write an email?”
  • Prompt: “Write a polite but firm email to my landlord. I want to announce a 15 percent rent reduction because the kitchen heating has been broken for three weeks. The tone should be factual and professional, about 150 words.”

The first version gives you a general explanation about email writing. The second delivers a finished email you can use right away. Same effort typing it in, completely different result.

Why Wording Changes Everything

Let's dig deeper into this, because this point is really crucial. The way you word your instruction doesn't just change the result a little. It changes it fundamentally.

Here are three before-and-after examples from different areas of life:

At the office:

  • Before: “Summarize this text.”
  • After: “Summarize the following project report in no more than five bullet points. Focus on the key results and open tasks. The summary is for my boss, who only has two minutes.”

The first version gives you a summary, but you don't know if it hits the right points. The second produces exactly what you need for your next meeting.

In a trade or craft business:

  • Before: “Create a quote.”
  • After: “Create a quote for a client who wants to renovate their bathroom. The space is about twelve square meters. New tiles and updated plumbing are needed. The tone should be professional but friendly. Please include line items, estimated costs, and a note about the warranty.”

A tiler who needs to send out quotes after a long workday saves a solid half hour with the second prompt. And the quote sounds more professional than anything they'd manage to type while exhausted.

In everyday life:

  • Before: “Help me study.”
  • After: “I'm preparing for my driver's theory test and I'm struggling with right-of-way rules. Explain the most important rules with simple examples. Then give me five practice questions I can answer.”

The first version gives you generic study tips. The second turns into your personal driving instructor. Same AI, same technology, completely different result. Just because of the wording.

The point is: the AI wants to help you. It just needs the right information to do so. And you provide that through your prompt.

The Core Principle: Clarity Beats Length

Now you might be thinking: “So I always need to write super long prompts?” No, absolutely not. Here's a principle that will help you from this moment on: Clarity beats length.

A short but clear prompt is better than a long but rambling one. It's not about typing as much text as possible. It's about providing the right information.

Think of it like a GPS. You could say: “Drive sort of south, somewhere near Chicago, there's this nice coffee shop, I think it's downtown somewhere.” Or you just type in the address: “200 Michigan Ave, Chicago.” The second way is shorter and still gets you exactly where you need to go.

The same applies to prompts. Instead of writing:

“I kind of need some sort of text, it's about marketing, for social media, like Instagram, and it should be engaging somehow, not too long but not too short either, maybe with emojis or maybe not, I'm not sure...”

You write:

“Write an Instagram post for a small flower shop. Topic: Spring bouquets are here. Casual, cheerful tone, max 150 words, with three fitting emojis and a call-to-action.”

The second prompt is shorter but much clearer. It contains everything the AI needs: Who (flower shop), for whom (Instagram followers), what (spring bouquets), how (casual, cheerful), and in what format (150 words, emojis, call-to-action).

Remember these five questions as your compass for every prompt:

  • What should the AI do? (The task)
  • For whom is the result? (The audience)
  • How should it sound? (The tone)
  • How much should it be? (The scope)
  • In what form should it come? (The format)

You don't always need to answer all five questions. But the more of them your prompt covers, the better the result will be. Over time, it becomes second nature.

From Wish to Precise Instruction

Let's get practical. How do you turn a vague wish into a precise instruction? Let's walk through a concrete example step by step.

Your wish: “I want to write a letter.”

That's a start, but the AI needs more. Let's go through the five questions:

  • What? A complaint letter to my internet provider.
  • For whom? For customer service, so professional and factual.
  • How? Polite but firm. Facts, not emotions.
  • How much? One page, no longer.
  • In what form? Classic letter format with date, subject line, greeting, and sign-off.

This becomes the following prompt:

“Write a complaint letter to my internet provider. For the past two weeks, my internet has been dropping out for several hours daily, even though I'm paying the full price. I want a credit and a quick resolution. The tone should be polite but firm. Please use a classic letter format with subject line and sign-off. Maximum one page.”

See how the vague “I want to write a letter” has become a clear instruction? The AI now knows exactly what to do and delivers a result you can use right away.

Another example from professional life: A carpenter thinks “I need something for my website.” That's the wish. The prompt could be: “Write a welcome text for the website of a carpentry workshop in the Portland area. The target audience is homeowners looking for custom furniture or kitchens. The tone should sound authentic and craft-oriented, no marketing buzzwords. About 200 words with one paragraph about experience (25 years) and one about philosophy (quality over quantity).”

From “I need something” to a text that actually fits. The difference is thirty seconds of thought and a few extra details in your prompt.

The beautiful thing: you don't have to be perfect at this. Even a prompt that answers just two or three of the five questions is far better than a vague question. And the more you practice, the more natural it becomes. Like riding a bike: at first you think about every move, and eventually you just do it.

Your Exercise: Three Variants of “Help Me Cook”

Now it's time to get hands-on. In this exercise, you'll experience firsthand how different the results can be when the same wish gets phrased in different ways.

Step 1: Open the prompt generator
Go to optiprompt.io and open the prompt generator.

Step 2: Enter your wish
Type into the input field: “Help me cook.” Select the category LLM.

Step 3: Test all three variants
Generate the structured, compact, and creative variants one after another. Copy each generated prompt and paste it into the AI tool of your choice, for example ChatGPT or Claude.

Step 4: Compare the results
Notice how differently the AI responds to the three variants:

  • The structured variant will likely ask about your preferences, ingredients, and dietary restrictions before delivering a detailed cooking plan.
  • The compact variant gives you quick, practical recipe suggestions without detours.
  • The creative variant might surprise you with unusual recipe ideas or a playful approach to cooking.

Step 5: Reflect
Which variant did you like best? Which result was most useful to you? There's no right answer. The goal is to develop a feel for how differently the same basic idea can be executed.

You'll see: from the simple “Help me cook,” the prompt generator creates three completely different instructions. And each one delivers a different but useful result. That's the power of a good prompt.

Conclusion: The Prompt Is Your Key

You now know what a prompt is: your instruction to the AI. You understand the difference between a vague question and a precise prompt. You've seen how wording fundamentally changes the result. And you've learned a simple principle that will improve every prompt you write from now on: clarity beats length.

The most important takeaway: you don't need to be a prompt expert to get good results. All it takes is a brief moment before each input to consider: What do I want, for whom, and in what form? Those thirty seconds of thought make the difference between “meh” and “that's exactly what I wanted.”

In the next article, “The Four Building Blocks of a Good Prompt,” we'll look at a system you can use to structure any prompt. Simple, logical, and immediately applicable.

Until then: try out the prompt generator. Take everyday wishes, turn them into prompts, and see what happens. You can't break anything. Every attempt is a step forward.

The art of good instructions begins with a single sentence. And you've just learned how to write it.

Autor

Sebastian Rydz

Das OptiPrompt Team teilt Wissen und Best Practices rund um KI und Prompt Engineering, um dir zu helfen, bessere Ergebnisse mit KI-Modellen zu erzielen.

Bereit, deine Prompts zu optimieren?

Erstelle mit OptiPrompt professionelle Prompts in Sekunden – kostenlos starten.