Your Teaching AI Toolkit
Prompts You'll Use Every Day
From Experimenting to Systematic
Over the last five chapters, you have used AI for lesson planning, assessment, personalization, and analytics. But if every interaction starts from scratch — typing a new prompt each time, remembering what worked last time, forgetting the phrasing that produced the best results — you are wasting effort. This chapter helps you build a reusable, organized toolkit of prompts and templates that becomes more valuable the longer you use it.
Think of it like your lesson plan binder. A first-year teacher creates every plan from scratch. A veteran teacher has a filing system of plans that they refine each year. Your AI prompt library is that filing system.
The Prompt Library: Structure and Organization
Look at data/prompt-library-education.json for a starter library of 40 prompts organized into six categories. Here is the structure:
| Category | Number of Prompts | When You Use Them |
|---|---|---|
| Lesson Planning | 8 | Sunday evening prep, unit planning, PLC meetings |
| Assessment | 8 | Creating quizzes, tests, rubrics, exit tickets |
| Differentiation | 6 | Before class, when creating tiered materials, IEP accommodations |
| Family Communication | 8 | After assessments, conferences, concern situations |
| Student Feedback | 5 | After grading assignments, progress reviews, report cards |
| Admin & Reporting | 5 | Quarter reports, data analysis, MTSS documentation, meeting prep |
Each prompt in the library has four parts:
Example: The "Weekly Quiz" Prompt
Name: Weekly Quiz Generator
Template:
Create a 10-question quiz for [GRADE] [SUBJECT], covering
[TOPIC/UNIT], aligned to [STANDARDS]. Distribution:
- 4 questions at Remember/Understand level (2 MCQ, 2 fill-in-the-blank)
- 4 questions at Apply/Analyze level (short answer, 2-3 points each)
- 2 questions at Evaluate/Create level (extended response, 5 points each)
Total points: 30. Time: 25 minutes.
Include an answer key with point values and acceptable response variations.
Language level: appropriate for [AGE]-year-old students (Lexile [RANGE]).Variables: GRADE (6-12), SUBJECT, TOPIC, STANDARDS (CCSS/NGSS/state), AGE, LEXILE RANGE
Notes: For AP courses, add "Include one AP-style free response question." For UK GCSEs, specify "mark scheme" instead of "answer key" and use GCSE command words (describe, explain, evaluate, assess). For Australian curricula, reference the ACARA content descriptor code.
Family Communication Templates
Communication with families is one of the most time-consuming non-teaching tasks. AI can draft these messages, but the tone matters enormously. A message that sounds robotic or generic does more harm than good.
Look at data/parent-communication.json for 15 templates covering common scenarios. Here are the categories:
| Scenario | Tone | Key Elements |
|---|---|---|
| Good progress update | Warm, specific, encouraging | Name one specific achievement, suggest how to continue at home |
| Declining performance | Concerned but supportive, never blaming | State facts without judgment, ask for partnership, suggest one action |
| Behavioral concern | Factual, solution-oriented | Describe specific incidents (not "your child is disruptive"), propose next steps |
| Conference invitation | Professional, welcoming | Specific date/time, what will be discussed, reassure nervous families |
| Absence follow-up | Caring, not accusatory | Express concern for the child's wellbeing first, then mention academic catch-up |
| Achievement celebration | Enthusiastic, proud | Make families feel their support contributed, invite them to share the moment |
The Multilingual Challenge
In diverse school communities, families may speak Spanish, Mandarin, Arabic, Vietnamese, or dozens of other languages. AI can help, but quality varies by language:
For high-stakes communications (IEP meeting invitations, disciplinary notices), always use your district's official translation services. AI translations are fine for informal progress updates and celebration messages, but legal documents require certified translation. Most US districts are required to communicate in families' home languages under Title III of ESSA.
Progress Report and Report Card Comments
End-of-quarter or end-of-semester report card comments follow a predictable structure, which makes them perfect for AI assistance. Instead of writing 28 unique comments from scratch, use this workflow:
Step 1: Prepare Your Data
For each student, have ready: assessment scores (3-4 data points), assignment completion rate, attendance percentage, one strength, one area for growth.
Step 2: Batch Generate
Prompt:
I need report card comments for 5 students. For each, I will
give you: name, assessment scores, assignment completion rate,
attendance, one strength, one growth area.
Format each comment as: 2-3 sentences, mentioning one specific
achievement, one specific area to work on, and one actionable
suggestion for the next quarter. Tone: professional, encouraging,
specific (not generic praise). Use language appropriate for a
formal report card.
Student 1: Sophia. Scores: 78, 82, 85. Assignments: 90%.
Attendance: 95%. Strength: consistent improvement in written
analysis. Growth: needs to attempt higher-order thinking questions.
Student 2: Aiden. Scores: 65, 60, 58. Assignments: 45%.
Attendance: 80%. Strength: excellent contributions during lab
activities. Growth: written expression and assignment completion.
[continue for remaining students]Generate in batches of 5-10 students. Review each comment, personalize where needed (AI will not know that Sophia volunteers for every science demonstration or that Aiden's family just relocated from another state), and adjust the tone.
Step 3: Quality Check
Read every comment aloud. If it sounds like it could describe any student in any school, it is too generic. Add one detail that only you would know. This is the difference between "Sophia shows consistent improvement" (generic) and "Sophia's willingness to revise her analysis of the primary source documents — and her growth from a 3 to a 4 on the evidence rubric — was one of the highlights of this quarter" (specific, memorable, genuine).
Assignment Feedback at Scale
Grading 28 essays or lab reports is exhausting. AI can help you provide more detailed feedback in less time — but only if you use it as a drafting tool, not a replacement for reading student work.
The Feedback Workflow
Prompt:
"I am giving feedback on a Grade 9 student's argumentative essay on climate change policy. My notes: (1) Strong thesis statement with clear position, (2) Evidence is cited but not analyzed — needs to explain how each source supports the argument, (3) Counterargument paragraph is weak — just dismisses the other side. Turn these notes into a 4-sentence feedback comment that is encouraging, specific, and gives one actionable tip for the next essay."
This takes 30 seconds per student instead of 3 minutes — and the feedback is more structured and actionable than what most of us write when we are tired and have 20 more papers to grade.
Versioning and Testing Your Prompts
Why Versioning Matters
Your first version of a prompt is rarely the best. Over time, you will discover that adding "use real-world US/UK contexts" dramatically improves relevance, or that specifying a Lexile range produces better-calibrated worksheets than "use simple language."
Keep a simple version log:
| Prompt Name | Version | Date | Change Made | Result |
|---|---|---|---|---|
| Weekly Quiz Generator | v1 | Jan 2026 | Original | Questions were too easy |
| Weekly Quiz Generator | v2 | Jan 2026 | Added Bloom's distribution | Good difficulty balance |
| Weekly Quiz Generator | v3 | Feb 2026 | Added "include one AP-style question" | Better for honors sections |
| Weekly Quiz Generator | v4 | Mar 2026 | Added Lexile range specification | Better reading level match |
Testing Before Deploying
Before using an AI-generated worksheet or quiz with students, run this checklist:
Building Your Toolkit Over Time
Your AI prompt library should grow organically. Here is a realistic timeline:
| Month | Focus | Expected Library Size |
|---|---|---|
| Month 1 | Lesson planning + quiz generation | 5-8 prompts |
| Month 2 | Add differentiation + family communication | 12-15 prompts |
| Month 3 | Add feedback + report card comments | 18-22 prompts |
| Month 4-6 | Refine existing prompts, add subject-specific ones | 25-30 prompts |
| Month 6+ | Share with PLC/department, create team-level library | 30-40 prompts |
The most valuable thing you can do is share your library with colleagues. A math teacher's differentiation prompt can be adapted for science in 30 seconds. An ELA teacher's family communication template works for every subject. When a PLC (Professional Learning Community) or department shares a prompt library, everyone benefits. Some districts have started building shared prompt repositories in Google Drive or their LMS — if yours has not, you could be the one to start it.
Your Complete Toolkit Checklist
By the end of this course, your toolkit should include:
Start with the prompts in data/prompt-library-education.json, customize them for your standards, grade level, and subject, and build from there. Within a semester, you will have a toolkit that saves you hours every week — hours you can spend on the parts of teaching that no AI can replicate: building relationships, sparking curiosity, and knowing each of your students as individuals.
Key Takeaways
This is chapter 6 of AI for Educators (Global).
Get the full hands-on course — free during early access. Build the complete system. Your projects become your portfolio.
View course details