The useful exam-review job for AI is diagnostic. Instead of asking for answers, ask it to find what you cannot explain yet, mix concepts together, and force you to choose the right method under exam-like constraints.
Use this when
- You have a study guide but do not know what to do first.
- You recognize definitions but miss application questions.
- You need a 20, 30, or 60 minute review plan.
- You want common traps before the exam.
Copyable starter prompt
Use this version when you already have a study guide, topic list, or set of old mistakes:
Act as an exam review coach. Use only the notes, review guide, and mistakes I paste below. Rank the topics by likely exam value, prerequisite importance, and my uncertainty. Build a 30-minute active-recall review plan. Create mixed practice questions, include common traps, and keep the answer key separate until after I attempt the questions.
What to paste into the prompt
Include the exam format, topic list, lecture titles, old quiz mistakes, formulas, and any professor emphasis you remember. The more specific the input is, the less generic the review plan becomes. If the course rewards application over memorization, say that explicitly.
What a useful result looks like
The output should start with a ranked topic list, not a summary. It should include active-recall blocks, application questions, traps, and a short fallback plan for the highest-yield topics if you run out of time.