AI prompting

Practical recommendation and reusable prompting resources developed to help faculty and staff prompt AI thoughtfully and effectively.

Example

“You are my co-designer for a drawing assignment. Help me edit this 30-minute lesson plan introducing perspective drawing to beginners. The students are community college learners with mixed experience levels. Output: a student facing step-by-step plan, that adheres to the provided learning objectives, and a 5-minute warm-up. Provide two options for me to choose from. Tone: see if you can bring some humor in to the assignment. Please use the example pdf provided.”

The Prompt Framework

(6 Key Components)

Prompts work best when they include these six components. You don’t always need all six, but the more clear you are, the more useful the response will be. 

Copy/Paste Prompt Template

Use this anytime you want stronger outputs:

Role: You are a…
Task: I need you to…
Context: This is for… (course/audience/goal)
Output format: Give me… (bullets/table/steps/script/rubric)
Tone: Keep it…
Example: Here’s an example of what I mean…

  • Tell the AI what role to take on.
    Examples: “You are an instructional designer…” or “Act as a writing coach…” 

  • Say exactly what you want it to do.
    Examples: “Create…” “Revise…” “Generate…” “Summarize…” 

  • Share the context so the response fits your situation.
    Include audience, course level, goals, time constraints, or assignment purpose.

  • Request the structure you want.
    Examples: bullet list, table, step-by-step sequence, rubric, script, checklist.

  • Describe the tone you need.
    Examples: supportive, direct, friendly, professional, plain language. 

  • Include a model when possible.
    Even a quick sample helps the AI match your style and expectations.

Improve Prompt Results

If the output feels “close but not quite,” try one of these adjustments:

  • “Give me a short version, a medium version, and a detailed version.”

  • “Before answering, list any assumptions you’re making.”

  • “Make this work for a 20-minute class.”
    “Keep the reading level around 8th grade.”

  • “Create a rubric to assess this work fairly and clearly.”

  • “Provide an accessible version with simple language and clear steps.”

Responsible & Inclusive AI Use

AI can support your work, but it should be used thoughtfully and try to keep it as workflow in an academic setting

When using AI for teaching and learning, keep these priorities in mind:

  • Avoid bias and stereotypes in examples, assumptions, and language 

  • Design for accessibility so materials are usable by more learners 

  • Consider cultural and social perspectives when generating content 

  • Use Universal Design for Learning (UDL) to support varied student needs 

  • Protect privacy (avoid entering personal student information into public tools)

  • Stay transparent about expectations for student use and authorship

  • Be specific with what you want the AI to search and also avoid

  • Question AI? if something does not sound right or is missing information

  • Remember that the AI is trying to give you more that what you asked for, so give it limits

Recursive Prompting

A recursive language model is a language model that feeds its own output back into itself—repeatedly—so it can refine, extend, or self-correct its response over multiple “passes,” instead of producing a single one-shot answer.

  • You are my thinking partner and studio assistant. Your job is to help me produce strong, usable work through short rounds of revision—like a draft, critique, and repaint.

  • Help me create or improve:
    [Write what you’re making: lesson plan / assignment / artist statement / rubric / critique / email / workshop outline / grant text]

  • Audience: [students / faculty / public / committee]
    Level: [beginner / intermediate / advanced]
    Format: [1 page / slides / spoken script / handout]
    Tone: [warm / direct / academic / casual]
    Constraints: [time limit, materials, policies, accessibility needs]

  • Paste your draft, notes, or messy idea—even if incomplete

  • You will complete the work in 4 passes.
    Each pass must build on the last. Don’t restart from scratch unless I ask.

  • Create a complete first version that is structured and ready to use.
    If information is missing, make reasonable assumptions and keep them simple.

  • Critique the draft using these quality checks:

    Clarity: Is it easy to follow the first time reading?
    Specificity: Are there concrete steps, examples, and outcomes?
    Flow: Does it move smoothly from start → middle → finish?
    Engagement: Does it invite participation instead of passive compliance?
    Equity + access: Can different learners succeed without extra barriers?
    Reality check: Does it fit the time, tools, and classroom conditions?

Next

Case Studies