Short answer: specificity is everything. The gap between a mediocre AI output and an excellent one is almost never the tool β€” it's the prompt. "Help me plan a maths lesson" produces generic filler. "Create 8 fluency questions and 2 reasoning questions on column addition for Year 3, using the method where we exchange ten ones for one ten, with an answer key" produces something you can nearly use as-is.

The four details that transform output

  • Year group and level β€” pitches vocabulary and difficulty correctly
  • Format β€” "as a table," "with an answer key," "as three tiered versions"
  • Constraints β€” time limits, materials available, word counts
  • The method you've actually taught β€” stops AI introducing an approach your class hasn't met

Build a library instead of starting fresh

When a prompt works well, save it. A personal prompt library β€” a simple shared doc, organised by subject or task β€” compounds in value over a term and is worth sharing across your year group. Our Prompt Library gives you tested starting points by role, and the subject-specific prompts guide goes deeper.