How to Write a Prompt That Actually Helps You Teach Better
Most educators who've tried AI tools for their teaching have had the same experience at least once: you type something in, you get something generic back, you spend twenty minutes fixing it, and you wonder whether it was worth the effort at all.
The tool wasn't the problem. The prompt was.
Prompting well is a skill, and like most skills, it looks simple from the outside but has real depth once you get into it. The good news is that a few consistent habits will get you most of the way there, and they're not complicated once you know what you're doing.
Why generic prompts produce generic outputs
AI language models work by completing a pattern. If your prompt is generic, say "write a lesson plan on photosynthesis," the model has very little to go on, so it produces the most statistically average response it can. It'll cover what most lesson plans on photosynthesis cover, at a difficulty level that matches most lesson plans on photosynthesis, in a format that resembles most lesson plans on photosynthesis.
The output isn't wrong. It's just not yours.
The more specific your prompt, the more specific your output. Context is the raw material that turns a generic AI response into something useful for your specific class, on your specific day, with your specific constraints.
The four things every good teaching prompt includes
1. Who your students are Not "Year 10 students." That's too broad. "Year 10 students who have covered cell biology but haven't started genetics yet, mixed ability class, about a third have English as a second language." The more accurate your description, the more appropriate the output.
2. What you're trying to achieve Not "a lesson on X," but what specifically you want students to be able to do by the end. "I want students to be able to explain the difference between mitosis and meiosis in their own words, not just recite the steps." Learning outcomes, not topics, drive useful outputs.
3. Your constraints Time, resources, setting, what students already know, what they struggled with last session. "45 minutes, no lab access, students found the diagrams confusing last week." Constraints are often where the most useful specificity lives.
4. The format you want Don't make the AI guess. "Give me a lesson plan in three sections: warm-up (10 min), main activity (25 min), and exit ticket (10 min)" is more useful than "write a lesson plan." If you have a template your institution uses, paste it in and ask the AI to fill it.
Examples by use case
Lesson planning
Weak prompt: "Write a lesson plan on persuasive writing."
Stronger prompt: "Write a 60-minute lesson plan on persuasive writing for a Year 9 class. They've already covered paragraph structure and can write a basic argument, but they struggle with using evidence effectively. I want them to leave being able to choose evidence that actually supports a claim, not just any evidence. Format: hook (5 min), direct instruction (15 min), guided practice (20 min), independent practice (15 min), reflection (5 min). No group work today — they'll be doing this individually."
Assessment questions
Weak prompt: "Write some questions about the French Revolution."
Stronger prompt: "Write five short-answer questions about the causes of the French Revolution for a Year 11 history class. One question should test recall, two should test analysis (asking students to explain why or how), and two should ask students to evaluate, meaning to make a judgement and support it with evidence. Each question should be answerable in 150–200 words."
Feedback on student work
Weak prompt: "Write feedback for this student essay."
Stronger prompt: "Write feedback for this essay from a Year 12 student. Focus on: (1) the strength of the thesis statement, (2) how well the evidence supports the argument, (3) one specific suggestion for the conclusion. Keep the tone encouraging but honest. The student is capable of A-level work but tends to rush the ending. Maximum 150 words."
The upgrade: from good prompt to built-in workflow
The natural endpoint of finding prompts that work consistently is making them permanent. Once you've figured out the prompt that produces a good rubric in your format, for your grading scale, for your discipline, you shouldn't have to rebuild it every time.
This is the idea behind Custom Skills in Noodle Factory: once you've identified the workflow that works, it gets embedded directly into the platform, already wired to your institution's templates, rubric language, and standards, so every output is already formatted correctly. The step from "good prompt" to "built-in workflow that anyone in my department can use" is what moves AI from a personal productivity tool to something with institutional value.
Start here
Pick one thing you need to prepare this week: a lesson plan, a rubric, a set of assessment questions. Write a prompt using the four elements above. Note how long it takes and how useful the output is.
Then try a second draft of the same prompt with more specificity. Compare the outputs.
That iteration (testing, comparing, refining) is how good prompting develops. It's not a course. It's just practice, with something real at stake.
Noodle Factory builds AI tutoring tools for universities, polytechnics, and K-12 schools across Southeast Asia. Learn more at noodlefactory.ai.
