Imagine you’re designing a course on climate change. You ask an AI to generate a quiz. It gives you 20 multiple-choice questions-none of them ask students to think critically. They all just test memorized facts. You could use this, but you know your students need more. You rewrite your prompt: "Generate five open-ended questions that ask students to compare historical climate policies in New Zealand and Canada, and explain how each policy reflects cultural values. Include one question that asks students to propose a new policy based on these comparisons." Suddenly, the output changes. The questions spark discussion. Students start researching. They care.
This isn’t magic. It’s prompt engineering-and it’s becoming as essential for course designers as lesson plans once were.
Why Prompt Engineering Matters in Education
AI tools like ChatGPT, Gemini, and Claude aren’t replacing teachers. They’re becoming teaching assistants that work 24/7. But they don’t know your goals unless you tell them clearly. A vague prompt like "Make me a lesson on photosynthesis" gives you generic, textbook-style content. A well-crafted prompt gives you something tailored: culturally relevant, aligned with your curriculum, and designed to challenge students at the right level.
Teachers in Auckland high schools are already using AI to generate case studies for economics classes. University instructors in Wellington are using it to create peer-review templates for writing courses. The difference? The prompts they used.
If you’re not learning how to talk to AI in a way that gets you useful, accurate, and pedagogically sound results, you’re leaving valuable support on the table.
What Prompt Engineering Actually Means for Instructors
Prompt engineering isn’t about using fancy words or coding. It’s about clarity, context, and control.
Think of it like giving directions to a new driver. If you say, "Go to the store," they might end up at the wrong one. But if you say, "Drive to the Countdown on Karangahape Road, next to the library, and park near the back entrance," you get exactly what you need.
Same with AI. Here’s what works:
- Define the role: "You are a high school biology teacher with 15 years of experience teaching NZ Curriculum Level 3."
- Specify the output format: "Create a 15-minute flipped classroom script with three discussion prompts and one real-world example from the Tasman Sea ecosystem."
- Set constraints: "Avoid jargon. Use simple language. Include one analogy that a 16-year-old would understand."
- Give examples: "Like this one: ‘Photosynthesis is like a kitchen where sunlight is the stove, CO2 is the flour, and water is the egg.’"
These aren’t tricks. They’re teaching techniques adapted for machines.
Five Prompt Templates Every Course Designer Should Know
You don’t need to invent prompts from scratch. Start with these proven templates-adjust them for your subject, level, and goals.
- Lesson Plan Builder: "Act as a curriculum designer for [grade/level]. Create a 45-minute lesson on [topic] using [teaching method]. Include: one engaging hook, two formative assessment questions, one group activity, and a real-world connection. Use [local context if applicable]."
- Quiz Generator: "Generate 10 multiple-choice and 3 short-answer questions on [topic]. Make sure 4 questions require analysis, not recall. Include an answer key with brief explanations. Avoid trick questions."
- Feedback Generator: "You are a writing tutor. Review this student paragraph: [paste text]. Give feedback that focuses on structure and clarity, not grammar. Suggest one improvement and ask one question to push their thinking. Keep it encouraging."
- Case Study Creator: "Create a 300-word case study about [real-world problem] in [industry/location]. Include: a protagonist, a decision point, two conflicting viewpoints, and one hidden bias. End with an open-ended question for class discussion."
- Adaptation Prompt: "Take this existing lesson on [topic] and adapt it for students with learning differences. Simplify language, add visual cues, break tasks into smaller steps, and suggest one alternative assessment method."
Copy these. Test them. Tweak them. Keep what works.
Common Mistakes Course Designers Make
Even experienced instructors stumble with AI prompts. Here are the top five mistakes-and how to fix them.
- Mistake 1: "Make it better." - Too vague. AI doesn’t know what "better" means to you. Fix: Say exactly what you want: "Make it more interactive," or "Make it suitable for ESL learners."
- Mistake 2: Ignoring bias. AI can reinforce stereotypes. If you ask for "a nurse," it might default to female. If you ask for "a CEO," it might default to male. Fix: Add: "Ensure gender-neutral and culturally diverse examples."
- Mistake 3: Trusting everything. AI hallucinates. It makes up sources, dates, and studies. Fix: Always fact-check. Use prompts like: "Only use information from peer-reviewed sources published after 2020."
- Mistake 4: Using AI to replace human interaction. AI can generate discussion prompts, but it can’t moderate them. Fix: Use AI to prepare, not to replace. Your role as an instructor is still central.
- Mistake 5: Not iterating. The first output is rarely the best. Fix: Treat AI like a junior colleague. Ask for revisions: "Make this more challenging," or "Add a real student quote."
How to Train Your Team to Use Prompt Engineering
If you’re leading a department or teaching team, don’t assume everyone knows how to use AI well. Start small.
- Hold a 30-minute workshop: Show two versions of the same prompt-one bad, one good. Compare the outputs. People remember visuals.
- Create a shared prompt library in Google Docs. Label each by subject and purpose: "History - Debate Prompts," "Math - Real-World Problems."
- Assign a "Prompt Champion" in each team. This person tests new prompts, shares wins, and troubleshoots issues.
- Include prompt engineering in your professional development plan. Make it as standard as learning LMS tools.
At the University of Otago, a pilot program trained 40 instructors over six weeks. By the end, 87% reported saving 4-6 hours per week on lesson prep. The biggest win? Student engagement went up-not because the AI did more, but because the lessons were better designed.
When Not to Use AI
AI isn’t a cure-all. There are times when it’s the wrong tool.
- When you need deep emotional insight: AI can’t understand grief, anxiety, or motivation the way a human can.
- When assessing creativity: If a student writes a poem or designs a sculpture, AI can’t judge its emotional impact.
- When building trust: Students need to know their instructor is present. Don’t outsource feedback to AI for high-stakes assignments.
- When ethics are involved: Don’t let AI decide who gets extra help or who’s "at risk." That’s your responsibility.
AI is a tool. Not a teacher. Not a decision-maker. A helper.
Next Steps: Start Today
You don’t need to become an expert overnight. Pick one task you do every week-maybe creating a quiz, writing feedback, or designing a discussion-and try rewriting your prompt using one of the templates above.
Test it. Compare the output. Ask yourself: Is this better? More engaging? More aligned with my goals?
If yes, keep it. If not, tweak it. Try again.
Over time, you’ll build a personal library of prompts that work for your style, your students, and your subject. That’s the real power of prompt engineering-it turns AI from a guessing game into a precision tool.
The future of education isn’t about who uses AI the most. It’s about who uses it the best.
Do I need to know how to code to use prompt engineering?
No. Prompt engineering doesn’t require coding. It’s about clear communication. You’re learning how to ask better questions, not how to write software. Most instructors use simple text prompts in tools like ChatGPT, Gemini, or Copilot without any technical skills.
Can AI create entire courses for me?
AI can generate content-lessons, quizzes, slides-but it can’t design a course that meets your students’ needs, fits your teaching philosophy, or responds to real-time classroom dynamics. You’re still the designer. AI is your assistant. Use it to save time, not to outsource your judgment.
How do I make sure AI doesn’t give me outdated or wrong information?
Always verify. Use prompts that ask for sources: "Cite peer-reviewed studies from 2020 onward." Check dates, names, and statistics. AI often hallucinates facts. Treat every output like a draft that needs your review.
Is prompt engineering just for universities?
No. It works at every level-from primary school to adult education. A kindergarten teacher can use it to create simple story prompts. A vocational trainer can use it to generate workplace scenario cards. The key is tailoring the prompt to your audience and goals.
What’s the biggest benefit of using prompt engineering in teaching?
Time. Instructors spend hours creating materials. With good prompts, you can generate high-quality drafts in minutes. That time gets reinvested in what matters most: talking to students, giving feedback, and building relationships.
If you’re designing courses today, you’re already working with AI. The question isn’t whether to use it-it’s whether you’ll use it well.