Teaching · Practical guide
Build a Study Plan with AI That Changes When Evidence Changes
A personalised plan should reflect your available time and the work you can currently do. It should not invent a learning style, diagnose a difficulty or promise a grade. This lesson builds a small plan around evidence and shows how to revise it after practice.
Collect useful planning information
Write down the assessment date, official topic list, permitted resources and realistic study periods. Describe what you can do with examples: “I can identify the numerator, but I often choose the wrong denominator.” That is more useful than a broad label such as “bad at maths.” Use non-sensitive descriptions; a planning assistant does not need your student identifier, medical history or private marks spreadsheet.
Choose a short baseline task for each important skill. Attempt it before viewing the solution. Record whether you answered independently, needed a hint or needed the full method. These categories describe this attempt, not your fixed ability. A confident self-rating is worth noting, but it should not replace a sample of actual work.
Treat the plan as a proposal. Confirm that the source list is real, the scheduled topics belong to your course and the activities fit the time available. A chatbot can invent chapters, assign inaccessible resources or place impossible workloads into a neat table. A visually impressive schedule still needs arithmetic and practical checks.
Plan the action, not just the topic
“Study percentages for an hour” does not explain what to do. A more usable block says: retrieve the formula without notes, solve two questions, compare with worked answers and record one correction. Each block should leave a visible product such as an annotated solution, a brief explanation or a corrected error log.
The Learning Scientists describe spaced practice as distributing learning over time and retrieval practice as trying to recall information rather than only rereading it. Use these principles modestly: revisit a topic on later days and attempt a question before checking notes. There is no single spacing schedule that guarantees success for every learner, subject and deadline.
Include time for feedback. If every scheduled minute is allocated to new material, errors can accumulate without correction. Reserve a small part of each session for checking solutions and deciding what to revisit. Include a buffer session for unexpected work or a topic that takes longer than expected; this makes the plan easier to adapt without pretending that lost time can be recovered magically.
Worked example: three sessions and a review rule
Fictional learner Noor has three 25-minute sessions this week and needs to calculate percentages and explain the denominator. A baseline attempt correctly identifies 15 attended out of 20 invited but reports 15%. This suggests a specific error to address: treating a count as a percentage. It does not establish a general mathematical difficulty.
Session 1: spend 5 minutes recalling the ratio-to-percentage method, 8 minutes following a worked example, 8 minutes solving two new examples and 4 minutes checking. For 15 of 20, the ratio is 15/20 = 0.75 and the percentage is 75%. Noor writes why the total invited belongs in the denominator.
Session 2, on a later day: use 4 minutes for recall, 10 minutes for two independent questions, 7 minutes to explain and repair errors, and 4 minutes to update the log. Session 3: use 5 minutes to retrieve the method, 12 minutes for a mixed task distinguishing counts, percentages and interpretations, and 8 minutes for checking and planning. Each session totals 25 minutes; the week totals 75 minutes.
Review rule: if either independent question uses the wrong denominator, start the next session with a contrast between the correct and incorrect ratios. If both are correct and explained, move to a new context while retaining a brief recall question. This rule uses performance evidence; it does not let the assistant automatically label the learner as strong or weak.
Adapt without rebuilding everything
After a session, record the actual task completed, an error or uncertainty, and the next action. “Finished page 4” reports activity; “used attendees instead of registered total as denominator” reports a useful learning issue. Give the assistant that short log rather than a long personal narrative.
If you miss a session, first protect the highest-priority learning goal. Reduce optional work or use the buffer. Do not simply double the next session when your availability has not changed. Ask the assistant to show what was postponed and why, so the revised schedule remains understandable.
Keep an independent check near the end of the plan. A learner who can reproduce the model answer immediately may still struggle with an unfamiliar case later. Use different numbers or a different context and explain the method without consulting the assistant first. If a topic remains unclear, record a precise question for a teacher or tutor.
Practice: design and revise a 20-minute planning exercise
Minutes 0–4: list three real study slots and one observable goal. Minutes 4–8: make a baseline question and attempt it. Minutes 8–14: draft three sessions with a product and checking time for each. Minutes 14–20: simulate one missed session and revise the plan while preserving the most important goal.
Example plan for a 60-minute weekly budget uses three 20-minute sessions: 4 minutes recall, 8 minutes practice, 5 minutes checking and 3 minutes recording next steps. If the second session is missed, Session 3 begins with the unresolved skill from Session 1 and postpones an optional extension. The revised week contains 40 minutes, not a fictional 60 minutes of completed work.
Example evidence log: “For 12 out of 16, I wrote 12% before checking; corrected to 75%; next step is to write subset/total before calculating.” A useful adjustment is to practise identifying both quantities in two fresh scenarios before performing the calculation.
Self-review: do scheduled minutes fit your actual availability? Does every block produce evidence you can inspect? Is there a specific rule for revisiting errors and a route to human help? Keep the plan simple enough to use; a smaller schedule you can follow is more useful than an elaborate plan you cannot maintain.
Reusable prompt
Select and copy this template, or customise your own using the free builder.
Create a study plan for this observable goal: [goal]. Official topics/resources: [approved list]. Available sessions: [days and minutes]. Baseline evidence: [non-sensitive attempt and error]. Include retrieval before notes, practice, feedback and a concrete output for each session. Show minute totals and one buffer or prioritisation rule. Explain how to adapt after a missed session or repeated error. Do not infer learning styles, diagnose difficulties or promise results.Open the prompt builder →
Pause · Think · Check
Test your understanding
Choose one answer per question, then check your reasoning. This is practice; scores are not saved.