Introduction
In the context of Generative AI, prompts refer to the information or questions entered into a Generative AI tool in order to get a response. The characteristics of the input influence the quality of the information received in the output. Prompt engineering is the process of refining a generative AI prompt to improve its accuracy and effectiveness.
Whether you are using Microsoft Copilot, ChatGPT, Gemini, Claude, or another tool, prompts are how you tell the AI tool what you need. There is no menu to navigate or button to click. You type a question or request in your own words, then continue the conversation by asking follow-up questions, refining your prompt, and building on the responses. This back-and-forth is when AI becomes most useful. AI is a collaborative partner that helps you think through ideas, solve problems, and adapt content to your needs.
Prompt Engineering
A great deal of advice has been generated on how to create prompts that yield information that is both relevant and accurate. This advice covers not only the content of the prompts, but also the process for developing and using them. Some advice has been codified into frameworks based on specific principles or steps (e.g., PARTS, CLEAR, CREATE, RICO, CARE); other advice has been offered in the form of lists of recommendations. Below is a summary of some of this advice.
Content
- Be specific: Add context and be clear and concise about what you want, based on clear goals and objectives (e.g., number of words, scope, level of detail)
- Assign a role: Ask AI to act like a certain type of person or someone in a certain role (e.g., high school teacher, organizational consultant, forensic psychologist).
- Use Do and Don’t: Give explicit instructions that set boundaries, define the limits of the task, and specify what you do and don’t want in the output. (e.g., do include statistics or don’t include sources prior to 2000).
- Specify structure and format: give precise instructions about the style of the output you want (e.g., outline, bullet points, summary, table, audio, image, or email)
- Give examples: Include an example of a sentence or paragraph that you would like the output to resemble; however, be sure not to infringe on copyrighted material.
- Consider tone and audience: Provide an idea of who the audience is and the tone you’d like to take (e.g., academic presentation, newspaper, popular culture magazine).
- Specify voice: Describe how you want the output to sound (e.g., enthusiastic, empathic, optimistic, or authoritative).
- Specify what evidence or sources to include: State explicitly if you want citations and references (and the format to use for them).
Process
- Build on previous prompts: Refine and iterate. Start with a basic question and add to it, changing words and making it more specific in order to create a conversation with follow-up questions.
- Correct mistakes and give feedback: Tell AI which parts of the output were useful, and which parts could be improved; point out mistakes.
- Try different phrasings and keywords if necessary: Rephrase prompts using synonyms.
- Ask AI to create prompts or indicate what else it needs: Start with the basic idea of what you want and ask the AI to expand on it by creating a prompt.
- Break down tasks into smaller steps: Enter these steps sequentially, one at a time (step by step). Use logical order, creating a logical flow of ideas in the prompt.
- Encourage creativity: Use phrases like “be creative” or “give me a new example”.
- Ask AI to explain its reasoning process: Ask for justification (“Explain your thought process” or “Why did you write that?”)
- Use simple language: Be brief and clear. Avoid confusing or ambiguous words that can have multiple meanings.
- Know when to start over: If you are beginning a new topic that doesn’t reference the previous exchanges, start a new chat instead of continuing the current conversation.
- Be reflective: Reflect on the AI output. Keep evaluating it and improving the prompts.
Examples of Prompts
The following 10 prompts are examples you can try. They are meant to help you get started by showing the kinds of tasks AI can support and how prompts can shape the response. These examples are a starting point, not a finished product. As you use them, continue the conversation by asking follow-up questions, refining your prompt, and building on the response. This back-and-forth helps you get more useful results and makes the interaction more collaborative.
| Task | AI Prompt | Why It’s Useful |
| Transform a lecture into a podcast | Turn this lecture outline into a conversational podcast script for students who prefer audio learning. | Supports multimodal learning and accessibility. |
| Make personal or interdisciplinary connections | Connect this education concept to playing baseball. | Helps faculty make personal or interdisciplinary connections and explore new teaching approaches. |
| Differentiate assignments | Design three versions of this assignment: beginner, intermediate, and advanced. | Supports differentiated learning by helping faculty adapt assignments for different skill levels. |
| Respond to student questions | Here’s a student question. Turn it into a 5-minute mini-lesson with examples and a quick quiz. | Helps faculty respond to common student questions with clear, structured support. |
| Create alternative assessment formats | Take this essay prompt and reimagine it as a video project, infographic, or podcast. | Encourages creative assessment and student choice. |
| Build rubrics from outcomes | Create a rubric for assessing collaboration skills in a group project. | Saves time and models backward design from learning goals. |
| Rewrite feedback for clarity | Rewrite this feedback in a way that’s encouraging and clear for a student needing support. | Supports compassionate, student-centered communication. |
| Anticipate student questions | Pretend you’re a student encountering this concept for the first time. What questions would you ask? | Helps faculty anticipate confusion and improve clarity. |
| Develop real-world case studies | Build a case study for my forensics course based on a [specific] recent event. | Helps connect course concepts to current, real-world contexts. |
| Support student reflection
|
Create three reflection questions that help students connect this unit to their personal goals or future careers. | Encourages reflection and helps students connect learning to their goals.
|
Sources
Harvard University Information Technology
Getting Started with Prompts
University of Michigan Generative AI Resources
Prompt Literacy in Academics
MIT Sloan School of Management Teaching & Learning Technologies
Effective Prompts for AI
Montana State University Library
Using Generative AI
University of North Carolina/Charlotte Office of OneIT
AI Helpful Hints: Creating Effective Prompts
Columbia University Digital Futures Institute
Tips and Tricks for Prompt Writing
Connecticut State Community College Northwestern Campus Library
Best Practices for Writing Effective AI Prompts
University of California/Davis Library
Generative Artificial Intelligence for Teaching, Research, and Learning
University of Pennsylvania Center for Excellence in Teaching, Learning, and Innovation
Prompting Strategies for Instructors
Google Cloud
Prompt Engineering: overview and guide
Cranfield University Library Services
Prompt Engineering
Sources for Prompting Frameworks
Monash University Student Academic Success
Creating effective prompts with AI
PARTS (Persona, Aim, Recipients, Theme, Structure)
Georgetown University Library
AI Resources: How to Create Prompts
CLEAR (Concise, Logical, Explicit, Adaptive, Reflective)
Lafayette College Technology Help
Best Practices for Prompting AI
CREATE (Character, Request, Examples, Adjustments, Type, Extras)
University of Pennsylvania Center for Excellence in Teaching, Learning, and Innovation
Prompting Strategies for Instructors
RICO (Role, Instruction, Context, Output)
University of Texas
Generative AI in Teaching and Learning: Prompt Design
CARE (Craft, Authenticate, Reflect, Evaluate)