Teachers are leaving the profession in huge numbers and honestly, who can blame them. Not because they stopped loving teaching but because the job has become something else entirely. Something buried under admin, paperwork, differentiation requirements, parent emails, and reporting cycles that have very little to do with why anyone became a teacher in the first place.

A 2023 RAND study found that 77 percent of teachers in the United States reported frequent job-related stress. The biggest culprit? Administrative tasks. On average, teachers spend 11 hours per week outside of teaching hours on things like lesson planning, grading, and communication. Eleven unpaid hours. Every week.

Here is the thing though. That number is starting to change for teachers who have figured out how to use AI properly. Not vaguely. Not occasionally. Properly, with specific frameworks that produce genuinely useful outputs. They are getting those hours back. Some of them are getting even more.

 

What Prompt Engineering Actually Is

Most people try an AI tool, type something vague, get a generic response, and conclude the technology is overhyped. The problem is almost never the AI. It is the prompt.

Prompt engineering is just learning how to talk to AI in a way that gets you something useful rather than something generic. Think of it like briefing a very capable new assistant. If you say “write something for my class” you will get something useless. If you say “write a 20-minute lesson plan for a Year 8 science class of 28 students on photosynthesis, including a 5-minute hook activity, three differentiated worksheets for foundation, core, and extension learners, and three formative assessment questions” you will get something you can actually use on Monday morning.

The difference is specificity. That is genuinely all it is.

 

The SPEC Framework for Lesson Planning

SPEC stands for Subject, Purpose, Experience Level, and Constraints. Fill in these four things before you write any lesson planning prompt and the output improves dramatically every time.

Subject is the exact topic, not just the subject area. Purpose is what students should be able to do by the end that they could not do at the beginning. Experience Level describes where your students are right now including common misconceptions. Constraints covers your time limit, class size, available resources, and any differentiation needs.

A well-structured SPEC prompt takes under two minutes to write and typically returns a lesson plan that would previously have taken 45 minutes to create from scratch. Teachers who have started using this consistently describe the same reaction every time. Genuine disbelief followed by genuine relief.

 

Grading Without Losing Your Weekend

Grading is where most of the weekly hours disappear, especially for written assignments. AI cannot and should not replace your professional judgment in final assessment. But it can be an excellent first-pass tool for formative feedback.

Give the AI your marking rubric. Give it a student response. Ask it to identify where the response falls on the rubric and suggest three specific, actionable feedback points. You review and personalise the suggestions before they reach the student. The whole process takes a fraction of the time writing everything from scratch requires.

For a class of 30 submitting essays, this approach consistently cuts marking time from four or five hours to under two. Always read the student work yourself first. AI feedback is a starting point, not a replacement for your professional read.

 

Parent Emails in Seconds

This one honestly surprised me when I first heard about it. Parent communication is one of the most emotionally draining parts of the week. Getting the tone right for a concern conversation, finding the right language for a behaviour incident without sounding accusatory, celebrating progress without overselling it.

AI drafts all of these beautifully if you give it three things: the context of the situation, the specific details, and the tone you need. A prompt that takes 30 seconds to write produces a first draft that takes another 60 seconds to personalise. The emails often end up better than what most people write from scratch because they start from a well-structured base and get shaped by the teacher’s specific knowledge of the family.

 

The Honest Bit About What AI Cannot Do

AI does not know your students. It cannot read the room, notice that a child seems upset today, adjust tone based on non-verbal cues, or build the trusted relationship that makes a student feel safe enough to ask an embarrassing question.

Everything AI produces for professional use needs a teacher’s eyes before it reaches students or parents. The goal is not to automate teaching. It is to reclaim the hours currently stolen by administrative work so teachers can spend more time on the human part of the job. The part that only they can do.

That is where the real work of teaching has always lived.

---