In February, the CSU announced a program that will offer members access to OpenAI’s ChatGPT Edu. As ChatGPT Edu rolls out and more staff, faculty, and students engage with this technology, education on how to use this tool effectively and communicate with AI becomes essential. This post will explore the basics of writing prompts and share tips and strategies for getting the most out of ChatGPT Edu.

What are prompts? Why is the way you write them important?

A prompt is what you type into a Large Language Model (LLM) like ChatGPT to tell it what you want. The prompts you enter into ChatGPT matter because they influence the quality and content of the response you will get. If the prompt entered is vague, misleading, or poorly structured, ChatGPT’s output will reflect that. On the other hand, clear and specific prompts help guide the model’s predictions toward more valuable outputs. 

Think about talking to ChatGPT like asking a librarian for a book. If you tell a librarian, “I need a book,” they’ll struggle to help. But if you say, “I’m researching ancient Roman architecture for a paper and need a book that covers structural techniques. Can you help me find books on this by next week?” You’ll get exactly what you need. The more detail and direction you provide, the better the AI (or librarian) can provide relevant information.

Prompt Writing Quick Tips

Explore these bite-sized strategies for improving prompts and consider how you might adjust how you talk to ChatGPT. You can incorporate more than one of these strategies in a single prompt.

• Be specific. The more details and context you provide, the more precise and valuable ChatGPT’s response will be. Avoid vague prompts that leave room for misinterpretation.
  • For example, “Summarize key strategies for engaging first-generation college students in online courses” is better than “How can I help students succeed?”

• Set the format and tone. If you want ChatGPT to provide answers in a specific style—like a bulleted list, paragraph, or email—include this in your prompt. The same goes for the length of response and tone.
  • For example, “Write a short, friendly email inviting faculty at a four-year public university to a professional development workshop about online learning.” is better than “Invite faculty to a workshop.”

• Specify the persona. Telling ChatGPT what role it should take helps it generate more relevant, expert-level responses, improving the depth and quality of its answers.
  • For example, “You are an instructional designer at a small four-year university skilled in course development. Walk me through the benefits of backward design when building online courses” is better than “What is backward design?”

• Use existing examples. If you want a response in a particular style, providing an example helps ChatGPT mimic it accurately.
  • For example, “Rewrite this syllabus introduction to be more engaging and student-friendly. Use tone and style similar to the following example: [insert text].” is better than “Make my syllabus introduction more engaging.”

• Ask ChatGPT to explain its reasoning. Asking for an explanation helps you verify the accuracy of responses and makes it easier to decide if the answers provided fit your needs. This is also a great way to learn new things when using ChatGPT to assist with unfamiliar tasks.
  • For example, “Recommend three strategies to improve student engagement in asynchronous courses and explain why each is effective” is better than “Give me three ways to engage students online.”

• Ask ChatGPT how to write a prompt. If you’re unsure how to phrase what you want, you can ask ChatGPT to generate a well-structured prompt. This can help refine your own prompt-writing skills, especially if you combine them with the above step and ask ChatGPT to explain the reasoning behind the steps in the prompt.
  • Example: “I want to analyze student survey responses for common themes. Help me write a clear, effective prompt I can use to ask ChatGPT for help with this. Ask me questions to make sure you have enough information.”

Ready for more? Try Prompt Chaining

What is it?

Instead of giving ChatGPT a single complex job, prompt chaining involves dividing one large task into several smaller and more manageable steps. This involves breaking down more extensive prompts into smaller, sequential prompts that you feed into ChatGPT one at a time. The output of one prompt serves as the input for the next, guiding the AI through a step-by-step structured process. 

Why does it work?

ChatGPT can better understand your intentions by breaking down the prompt into smaller and more manageable requests. This strategy also provides you, the human-in-the-loop, greater control at each process step. Prompt chaining makes it easier for you and the AI to understand the steps required to complete a request, provides more opportunities for you, the human user, to make adjustments, and reduces the risk of errors or misunderstandings.

What does it look like?

• Original prompt:
  • Write a clear, student-friendly FAQ for first-generation college students.
• Prompt-chaining:
  • Prompt 1: “Generate a list of 15 common questions first-generation college students might have about navigating college.”
  • Prompt 2 (using the question list from Prompt 1): “Remove the question about and add a question about . Then, group these questions into categories like academics, financial aid, and student life.”
  • Prompt 3 (using the groupings from Prompt 2): “Provide clear, student-friendly answers to each question. Keep responses under 100 words and use an encouraging tone.”
  • Prompt 4 (using the answers from Prompt 3): “Make the responses more conversational and inclusive [or anything else that you, the human, thinks would be a good fit], addressing common concerns with empathy. Tailor responses to a student body that [relevant information about student body that is important to include according to you, the human]”
  • Prompt 5 (using the refined response from Prompt 4): “Reformat the FAQ as a bulleted list with bolded questions.”

What Do You Think?

For anyone choosing to use AI tools like ChatGPT Edu, learning to craft effective prompts carefully is a new skill that is important to work on consistently. This post is meant to be a starting point for your journey in getting better outputs from AI—no single blog post can capture all the details of prompt engineering. Continue your education by testing the strategies outlined in this blog, reading more about prompt engineering (consider the references below), and researching strategies to fit your more advanced needs.

References

General Tips for Writing Prompts. (2024). Prompt Engineering Guide. https://www.promptingguide.ai/introduction/tips

Getting started with prompts for text-based Generative AI tools. (2023, August 30). Harvard University Information Technology. https://www.huit.harvard.edu/news/ai-prompts

Kuka, V. (2025, March 6). Introduction to Prompt Engineering. Learn Prompting. https://learnprompting.org/docs/basics/prompt_engineering

Prompt engineering: Six Strategies For Getting Better Results. (n.d.). OpenAI Platform. https://platform.openai.com/docs/guides/prompt-engineering#six-strategies-for-getting-better-results

What is prompt chaining? (2024, December 19). IBM. https://www.ibm.com/think/topics/prompt-chaining