AI-assisted learning · Daily practice

Prepare today’s practice from yesterday’s mistakes

Use actual answer history to decide what to practice. This example produces one new set targeting the most supported weak question type.

On this page
One goal for this article

Read answer history before making new questions of the same test format.

A concrete example

Illustrative example, not your study record

Illustration: 12 answers yesterday; four wrong kanji readings and two wrong grammar patterns. These are example numbers, not your history.

What a good result looks like

Illustrative result; verify against real data

Report the date, totals, and wrong answers by question type first. Explain why reading is the priority, then create eight new reading questions with answers and explanations. Verify the saved practice title and count.

A prompt you can use

Read my previous day’s answers in JLPT Master using Asia/Tokyo time. Report the totals and errors by question type. Make eight new questions of the most supported weak type, with different contexts, answers, and explanations for each option. Save today’s formal practice and report its actual title. If yesterday has no answers, explain that and inspect older history. Do not invent accuracy or sources.

Three practical steps

  1. Check which date and answers were actually read.

  2. Verify that new questions test the same skill in new contexts.

  3. Open Today’s Practice and inspect the saved questions and explanations.

Where to check the result

Find the new set under Today’s Practice. Chat text, drafts, and formal practice are different states.

View today’s practice →

What to watch for

A small number of answers suggests a direction, not a stable diagnosis of your weakest skill.