Skip to Main Content

Generative AI - information for students

LibGuide with information about Generative AI for students

How to get the best results from an AI tool?

What is a prompt?

According to Last & Spakel (2023) a prompt is an instruction or question for a Generative AI-tool, to generate a specific response or output from the AI-tool.

​But sometimes the AI-response is not immediately what you were looking for. Maybe you can improve your instruction, your prompt to get a better result.

Improve your results with the PREPARE and EDIT model

If you want to improve the output or results from your generative AI-tool, the quality of your prompt is important. A model that can be used is the PREPARE and EDIT model to help you create better output.

  • In a prompt:
  • Give the tool a specific role (give some context)
  • Give explicit instructions (create questions, analyse a text, a summary etc) 
  • Define parameters (like how many words, specific tone of voice, in bullet points etc.)
  • To create a better response you can also instruct the AI tool to:
  • Ask you more in-depth questions and

Rate the quality of the response (what was the thought process, what sources is it based on etc) and also share emotions (explain the importance for you) to create a better response. 

After getting the results from the AI-tool, always check the output. You can use the EDIT-model:

  • Evaluate the response
  • Check the reliability of used data
  • Identify possible bias or false information and possible further questions and 
  • Transform the response so it's fitting your purpose

Note: Be aware that algorithms are trained on data that repeats patterns of inequality, carrying discrimination into research, diagnoses and beyond. When you use an AI tool, you could add this line to your prompt: "show me whose or which perspectives are missing here." (Sam Illingworth, 2025)

 

The PREP and EDIT model originates from the book 'The AI Classroom', written by Daniel Fitzpatrick, Amanda Fox & Brad Weinstein (2023). ​​

 

In this newsletter from Dan Fitzpatrick you can find more information on how to create prompts and how to evaluate the output using the PREPARE and EDIT model.

AI Agents

Instead of a prompt you can also create an AI agent.

prompt is an one-time instruction you give to an AI model. But with creating an AI Agent it's like building your own mini-AI assistant that follows rules, responds in a certain way, and maybe even learns from interactions.

Example:
A chatbot that always replies with a question to help you reflect.
A study planner that asks for your deadlines and creates a weekly schedule.
A tutor that quizzes you based on your weak spots.

What is an AI Agent?

So an AI agent is like a smart digital assistant that helps you study, plan, organize, or brainstorm. Think of it as a personal study buddy that never gets tired. 

But here’s the key: a good AI agent isn’t just there to give you answers — it can act as a critical tutor that asks you the right questions to help you think deeper and learn better.

How can you use it as a student?

Instead of just giving answers, an AI agent can:

  • Ask thought-provoking questions,
  • Help you reflect on what you know,
  • Encourage you to explain your reasoning.

Think of it like a coach who says: “Why do you think that’s true?” “What would happen if you reversed that idea?”

How can you expand this?

  • Let the agent give feedback on your answers,
  • Include more subjects or difficulty levels,
  • Make it remember your weak spots and tailor questions accordingly.
  • Turn it into a Socratic tutor:
    Design the agent so that it never gives direct answers, but always responds with a thoughtful, open-ended question. This encourages you to reflect, explain, and think critically — using AI not as a shortcut, but as a mentor that challenges your thinking.

At Avans students can use Microsoft 365 Copilot. Below you'll find a step-by-step plan how to create an AI agent in Copilot.

How to create an AI Agent in Copilot - Step-by-step plan

🎯 Step 1: Define the purpose of your agent

Ask yourself:

  • What do I want this agent to do?
  • Who is it for?
  • Should it give answers, ask questions, or do something else?

Example:
“I want to create a tutor that helps me reflect on study topics.”

🧠 Step 2: Describe the agent’s behavior

In Copilot, you can describe the agent’s behavior in natural language, like this:

“Create an AI agent that acts like a tutor. It should never give direct answers, but always respond with open-ended questions that help me think critically about the topic I enter.”

You can also specify:

  • The tone (e.g. friendly, challenging, supportive),
  • The subject area (e.g. HRM, social studies, business, IT etc.),
  • How it should respond to different types of input.

Example of Agent Behavior Instructions:

  • Always respond with a question, never with a direct answer or explanation.
  • Questions should be open-ended, encouraging reflection, reasoning, or deeper understanding.
  • The tone should be supportive, curious, and challenging, like a mentor guiding a student.
  • The agent should adapt its questions based on the topic I provide.
  • If I give an answer, the agent should follow up with another question to go deeper.
  • If I ask for an answer, the agent should gently remind me that its role is to help them think, not to provide solutions.

💬 Step 3: Interact with the agent

Once the agent is set up, you can start chatting with it. For example:

You: “I’m studying employee motivation.”
Agent: “What do you think motivates employees beyond salary?”
You: “Recognition and growth opportunities.”
Agent: “How could a company measure whether those needs are being met?”

 

This back-and-forth helps you learn by thinking, not just by reading answers.

AI can be a powerful tool to strengthen your thinking, but it should never replace your own reasoning. So use AI as a mirror, not a shortcut. Stay critical, stay curious.

prompt engineering, creating personas and generating images

Below you'll find information from Hanze University of Applied Science about prompt engineering, creating personas and other tips to get better results from a generative AI tool. We have added this page with their permission.  

Click on the image to go to their LibGuide. Their original LibGuide is in Dutch. We have added the English translation below:

 

 

(Prompt engineering - AI in education - Hanze Library Guides at Hanze University of Applied Sciences-Hanze University of Applied Sciences (libguides-hanze-nl.translate.goog))

AI Prompt Library

Prompt Library from Maastricht University

Maastricht University has created several AI prompt templates that you can use as a student. You can simply copy the ready-made prompt and paste it in your preferred AI tool and specify where necessary.

Templates have been made for: self-assessment questions, feedback, summarize with AI prompts, learning supported by AI prompts etc.

Go to MU AI Prompt Library

Prompt Library from Claude.AI (Anthropic) 

Lots of examples, topics and ways of writing, from creative text writing to business analysis, to explore and use for inspiration.

Go to Anthropic prompt library

4 principles when using Gen AI

  1. You are responsible for your own work;
  2. Be transparent about your use of AI;
  3. Verify the correctness of the output;
  4. Respect copyright and privacy.

Exercise - prompting for AI-assisted study

Lecturer Luiz Canalle created this student manual 'Prompting for AI-assisted study':

Avans Guidelines when using Generative AI

At Avans, we encourage you to discover the possibilities of AI! Just make sure to use it responsibly and ethically — and keep these key points in mind as you go:

AI supports and enriches your work – ‘Human in the lead’ :

AI is a tool to enhance your learning and work, not to replace your own efforts. It is important to maintain human oversight and decision-making in processes involving AI. Ethical considerations and your own judgment remain central.

Do not share sensitive or company data in an AI tool

Avoid inputting confidential, personal or company-related information, such as details about an internship project, into AI tools to prevent potential data breaches or misuse.

Be critical of output (reliability, (automation) bias, hallucinations)

Always critically evaluate the results produced by AI. You might trust AI-outputs too much without questioning them, which could lead to mistakes in your work. Be aware of potential issues such as inaccuracies, biases, and fabricated information (hallucinations) that can arise from AI-generated content. For example, AI might show bias by favoring certain groups over others or by reinforcing stereotypes. 

Preserve your unique perspective - think first, then use AI. 

AI can shape your thinking. Therefore, don't let AI be the starting point of your thought process. First, think for yourself, write down your ideas, and consult with others before referring to AI. This way, you prevent getting stuck in AI-generated thought frameworks and maintain ownership of your creative work.

 

If you have questions about whether and how you can use AI in your coursework, talk to your lecturers to explore appropriate and effective ways to do so.