Lesson 1 of 6·7 min read

Prompt Engineering Basics ✍️

The difference between a mediocre and an excellent AI result almost never comes down to the model — it comes down to the prompt. Prompt engineering is the skill of giving AI models the right instructions. Master it, and you multiply your productivity.


🎯 What You'll Learn

  • What a prompt is and why formulation matters
  • The 4 elements of a professional prompt
  • Common anti-patterns that cause poor results
  • The iterative workflow for prompt optimization

What is a Prompt? 📝

📖 Definition: A prompt is the input you give an AI model — a question, an instruction, or a complex template. The prompt largely determines the quality, relevance, and usefulness of the AI's response.

The difference in practice is dramatic:

❌ Weak prompt:

"Write me something about marketing"

✅ Professional prompt:

"Create a 5-point checklist for planning a social media campaign for a B2B SaaS company with 50 employees. Target audience: IT decision makers in the DACH region. Tone: professional but approachable. Format: Numbered list with 2-3 sentences of explanation each."

The second example typically delivers a usable result on the first try. The first requires multiple follow-up questions.


The 4 Elements of a Good Prompt 🧩

ElementWhat It DoesExample
🎭 RoleDefines expertise and perspective"You are an experienced B2B SaaS marketing consultant"
📋 ContextProvides relevant background information"Our company has 50 employees and sells HR software"
🎯 TaskPrecisely describes what should be done"Create a campaign checklist with 5 points"
📐 FormatSpecifies the desired output structure"As a numbered list with 1-2 sentences each"

💡 Tip: Not every prompt needs all 4 elements. For simple questions, a clear task is enough. But the more complex the task, the more elements you should use.

Complete example with all 4 elements:

[Role] You are a Senior Data Analyst with experience in the
logistics industry.

[Context] Our company delivers 2,000 packages daily in the
DACH region. Average delivery time is currently 3.2 days.

[Task] Analyze the following delivery data from the past 6
months and identify the top 3 causes of delayed deliveries.

[Format] Present results as:
1. Executive Summary (3 sentences)
2. Causes table (Cause | Frequency | Recommendation)
3. Immediate action for each cause

Anti-Patterns: What to Avoid 🚫

Anti-PatternProblemBetter Alternative
"Make it good"No clear quality definitionSpecify concrete criteria
"Be creative"Too vague, unpredictable results"Create 3 unconventional variants"
Everything in one promptOverwhelms the modelBreak into steps (prompt chaining)
No length constraintResponse too long or too short"Max 200 words" or "5 bullet points"
Double negationConfusing for the modelPhrase positively

⚠️ Caution: The most common anti-pattern is missing context. The AI knows nothing about your area of work, industry, or target audience — you need to actively provide this information.


The Iterative Workflow 🔄

Prompt engineering is not a one-time action. The best results come from systematic refinement:

  1. First draft → Formulate prompt with the 4 elements
  2. Evaluate → Is the result useful? What's missing?
  3. Refine → Add missing details, set constraints
  4. Test → Check with different inputs
  5. Document → Save the successful prompt in the team library

🔑 Remember: A prompt that works well once works well every time. Invest the time in optimization — it pays off with every reuse.


📋 Summary

  • The quality of the prompt determines the quality of the result
  • Use the 4 elements: Role, Context, Task, Format
  • Avoid anti-patterns like missing context or vague instructions
  • Prompt engineering is iterative — refine, test, document
  • Build a team library of proven prompts

🎯 Exercise: Take a task you do regularly and write a prompt using all 4 elements. Test it and refine it through 3 iterations.


Next lesson: Role Prompting and System Prompts — how to assign expertise to the AI.

📝

Quiz

Question 1 of 4

Welche 4 Elemente gehören zu einem guten Prompt?