Ask an AI for a study plan and you will get a tidy timetable that assumes you have nothing else on. It looks organised, so you save it — and by week two it no longer matches your actual life, so you stop opening it. The problem is not the model. It is that it was never told what it needed to know.
Give it your constraints before your subjects
Most people open with the topic: "make me a revision plan for biology." The model has no idea how many hours you have, when your exam is, or what you already know, so it invents all three. Every number in the plan is a guess, and guesses are what make a plan collapse on contact with a real week.
Lead with the constraints instead. The exam date. How many evenings you genuinely have free, not how many you wish you had. Any fixed commitments — a shift, a sport, a commute. Whether you are starting from lecture notes or from nothing. A plan built around four real evenings beats one built around seven imaginary ones.
Work backwards from the exam, not forwards from today
Ask for the plan in reverse: what has to be true the week before the exam, then the week before that. Working forwards fills the early weeks with whatever is easiest and leaves the hard material stranded at the end, which is exactly where you have the least time and the most stress.
Backwards planning also exposes the honest answer early. If the plan cannot fit in the time you have, you find out in week one, while you can still cut scope or ask for help — not the night before.
Make every session a task, not a topic
"Tuesday: revise photosynthesis" is not a session, it is a heading. You will sit down, reread the chapter, feel vaguely productive, and retain very little. Ask the model to turn every block into something with a finish line.
| Instead of | Ask for |
|---|---|
| Revise photosynthesis | Answer 10 past-paper questions on photosynthesis without notes, then mark them |
| Read chapter 4 | Summarise chapter 4 into 8 flashcards, then test yourself on all 8 |
| Work on the essay | Write the introduction and pick three sources for the second section |
| Go over lecture notes | Explain last week's lecture out loud from memory, then check what you missed |
The difference matters because a task tells you when to stop. A topic does not, so it expands to fill the evening and still feels unfinished.
Build in a catch-up day you expect to use
A plan with no slack is a plan that breaks permanently the first time you miss a day. Ask for one empty block a week and treat it as part of the plan, not as failure. If you do not need it, you get an evening back. If you do, the plan absorbs the miss instead of unravelling.
Re-plan weekly instead of rewriting from scratch
At the end of each week, tell the model what actually happened — what you finished, what you skipped, what turned out harder than expected — and ask it to adjust the remaining weeks. This is the step almost nobody does, and it is the one that keeps a plan alive. You are not starting over; you are correcting course with real information.
A prompt to start from
Fill in the brackets and paste this. It is deliberately blunt about constraints, and it asks the model to push back before it plans:
"My [subject] exam is on [date]. Today is [date]. I have [number] free evenings a week, roughly [number] hours each, and these fixed commitments: [list]. Here is the syllabus: [paste]. Before you plan anything, ask me to rate how confident I am on each topic from 1 to 5. Then work backwards from the exam date and give me a week-by-week plan where every session is a specific task with a finish line, not a topic. Leave one empty catch-up block each week. If the syllabus does not fit in the time I have, tell me what to cut instead of shrinking every session."
When the plan breaks anyway
It will, and that is not a reason to abandon it. Missing two days is a scheduling problem with a scheduling fix: report what happened, ask for a revised plan, carry on. The plan is a tool, not a promise. The only version that fails completely is the one you stop updating.
Skip the setup — build your plan free
Our free AI planner asks for these constraints up front and builds the week-by-week plan for you. No account, no email required.
Build my study plan freeWant the prompts ready-made? The Student AI Prompt Pack collects 55 of them — essays, exam prep, recall, and focus — written the same way as the one above.