How to create good prompts

Prompting is the art of communicating effectively with a generative AI language model. When you write a prompt, you give GenAI an instruction, a question, or a context that the model uses to generate a response. The quality of GenAI’s output depends heavily on the quality of your input.

Think of it as assigning a task to a highly competent colleague who has never worked with you before: the more precise you describe the task, the context, and your expectations, the better the result will be.

As a student, the best prompt is the one that makes you think more – not less. Use generative AI as a patient learning partner, not as a shortcut. Your future career is built on understanding, not on the ability to get the right answers from a machine.

 

The five basic principles of good prompts

Before diving into specific techniques, here are five principles that apply regardless of which technique you use.

  • Be specific and clear
    Avoid vagueness. Write exactly what you want. “Write a draft for a welcome email to new employees” is better than “write an email."
  • Provide context
    Tell GenAI who you are, who the recipient is, and what the purpose is. Context makes the response relevant and precise.
  • Specify format and scope
    Do you want a short summary, an in depth explanation, a table, or a bullet list? State this explicitly.
  • Iterate and improve
    The first prompt is rarely perfect. Use the responses to refine your prompts. Prompting is a conversation, not a single command.
  • Give examples when possible
    Examples of the desired output are often more effective than long descriptions of what you want.

 

Generative AI as a learning partner

GenAI is an extraordinary learning tool – but only if you use it to strengthen your understanding, not replace it. The difference lies in how you prompt. The examples below show how you can consistently use GenAI to learn more and go deeper, rather than just getting an answer you can copy.

This guide walks through the most important techniques and provides concrete examples that you can adapt to your own needs.

Learning technique What you prompt What you learn
Socratic dialogue "Ask me questions instead of giving answers" Deep conceptual understanding
Reverse teaching "Evaluate my explanation of …” Identification of your own knowledge gaps
Error analysis "Find the error but don’t give me the solution" Error recognition and debugging
Concept exploration "Help me connect these concepts" Understanding of relationships
Exam simulation "Act as my examiner" Exam readiness and self evaluation
Theory to practice "Give me real world examples of..." Practical intuition

 

Instead of asking GenAI to explain something, you ask it to pose questions that guide you toward discovering the answer yourself. This method significantly strengthens deep understanding.

Learning example

Prompt:

I am studying Gibbs free energy in my thermodynamics course. Instead of explaining it, I would like you to use the Socratic method. Ask me one question at a time that leads me to understand on my own: - • What Gibbs free energy tells us - Why it is useful - When a reaction is spontaneous. Wait for my answer before asking the next question. If I am wrong, give me a hint instead of the answer.

Why this supports learning:

You are forced to think actively. You receive feedback on your mistakes without being given the answer. The deep understanding you build is far more durable than passive reading.

Explaining something to others is one of the most effective learning methods. Use GenAI as your “student” and explain concepts to it – then ask it to evaluate your explanation.

Learning example

Prompt:

I want to practice explaining things. I will give you my explanation of a concept, and you will evaluate it. Here is my explanation of the Reynolds number: “The Reynolds number is a dimensionless number that describes the relationship between inertial forces and viscous forces in a fluid. When it is high, inertial forces dominate and we get turbulent flow. When it is low, viscosity dominates and we get laminar flow.” Evaluate my explanation: 1. Is it academically correct? 2. What is missing? 3. What could be explained more clearly? 4. Give me a follow up question that tests whether I truly understand the concept. .

Learning Example

Prompt:

I am trying to understand convective heat transfer. I think of it as blowing on hot soup – the faster you blow, the faster it cools. Is my analogy correct? Where does it work, and where does it break down? Suggest a better analogy if mine is incomplete.

When you make mistakes, it is a golden opportunity for learning. Instead of asking GenAI for the correct answer, ask it to help you understand your mistake.

learning example

Prompt:

I tried to solve this energy balance problem, but my answer is incorrect. Here is my solution: [insert your solution here] Do NOT give me the correct answer yet. Instead: 1. Identify exactly where in my solution the error occurs 2. Explain the misunderstanding behind the error 3. Give me a hint so I can correct it myself 4. When I have tried again, will you check my new answer?

Why this supports learning:

Understanding why something is wrong is far more important than knowing the right answer. This method builds error recognition – a core competence for engineers.

learning example

Prompt:

My Python code for simulating a batch reactor produces incorrect results. Here is the code: [insert code] Instead of fixing the code for me: 1. Tell me which line(s) are problematic 2. Ask me questions about what I think the code does in those lines 3. Help me understand the concept behind the error 4. Let me try to fix it myself.

 

Use GenAI to explore concepts from multiple angles. This goes beyond simply getting a definition – the goal is deep, connected understanding.

learning example

Prompt:

Help me build a mental concept map of transport phenomena. Start with these three pillars: mass transfer, heat transfer, and momentum transport. For each pillar: - What is the fundamental driving force? - What is the analogous law (e.g., Fourier’s law)? - Give a concrete engineering example. Then show the mathematical parallels between them. Finally, ask me a question that tests whether I have understood the connections.

learning example

Prompt:

I am learning about chemical equilibrium. Help me explore the concept through “what if” scenarios. Use the Haber–Bosch process as the starting point. Ask me a series of “what if” questions, one at a time: - What if we change the temperature? - What if we change the pressure? - What if we remove product? Wait for my answer and gently correct me if I am wrong. Build on my answers with more challenging questions.

Generative AI can be an excellent exam partner – but again, the focus is on testing and strengthening your understanding, not on receiving ready made solutions.

learning example

Prompt:

Act as my examiner in the course Chemical Unit Operations. Start with a question of medium difficulty. Rate my answer on a scale from 1–10 and explain how I can improve. - If I score above 7, move on to a more difficult question. - If I score below 5, give me a hint and let me try again. After 5 questions: give me an overall evaluation and identify my knowledge gaps.

learning example

Prompt:

I am preparing for an exam and I have memorized Bernoulli’s equation: P₁ + ½ρv₁² + ρgh₁ = P₂ + ½ρv₂² + ρgh₂ But I am not sure I truly understand it. Test my understanding by: 1. Asking me to explain what each term represents physically (not just mathematically) 2. Giving me a situation where it does NOT apply and asking me to explain why 3. Giving me a practical scenario and asking me to set up the equation (without solving it) .

At a technical university, abstract concepts can feel far removed from reality. Use GenAI to build a bridge between theory and practice.

learning example

Prompt:

I have just learned about PID control in my process control course. I understand the mathematics, but I struggle to connect it to real life. Give me three concrete industrial examples of PID control. For each example: • What is the process? - What is the setpoint? - What happens if the P term is too large? - What happens if the I term is missing? - How would an operator experience poor tuning? Explain it so that a person without process experience can visualize it.

This page is inspired by Aalborg University’s page on generative AI for students. Generative AI has been used in the preparation of this page.