Good exit quiz writing is
a pattern-matching task.
Writing a well-designed exit quiz requires two things: understanding the concept being assessed, and knowing which question types produce actionable data at which cognitive level. The first is domain knowledge. The second is a learnable pattern. AI is good at the second — when given the domain knowledge, it can apply the pattern reliably.
The pattern is the 3-question check from C3/A2: one recall question, one understanding question, one application question. Each follows a design rule. These rules are fixed and transferable — AI can apply them to any subject area once given the learning objective and the relevant domain knowledge.
Confidence rating or content restate (vague data): 2 minutes.
AI exit quiz with the prompt below (actionable 3-question check with diagnostic key): 30 seconds generation + 90 seconds editing = 2 minutes total.
Same output quality as manual. Same actionability. 90% time reduction.
The exit quiz prompt that produces
actionable output every time.
The prompt has five components. The first two (learning objective and class profile) determine content accuracy and calibration. The third (question format specification) ensures the AI uses the 3-question check structure rather than defaulting to recall-only. The fourth (cognitive level specification) ensures each question operates at the right level. The fifth (diagnostic key) converts the quiz from a data collection instrument into a decision tool.
Details:
Learning objective: [students should be able to ___]
What was covered: [2–3 sentences summarising today's lesson content and examples used]
Class profile: [year group, what they already knew before this lesson]
Format required: 3 questions in this exact sequence:
Q1 — Recall: one correct answer, no inference required, tests retrieval of a specific fact or term from today's lesson
Q2 — Understanding: requires explanation of a mechanism or relationship; cannot be answered correctly using only memory
Q3 — Application: uses a novel context NOT covered in today's lesson; requires transferring today's concept to a new situation
For each question include: the question text, the correct answer (2–3 sentences), the most likely wrong answer, and what misconception that wrong answer reveals.
Q3 constraint: the application scenario must be genuinely different from all examples used in today's lesson.
Why the prompt must name the level —
not just the format.
Without explicit cognitive level specification, AI defaults to recall and comprehension. Even when you ask for an “application question,” AI often produces a comprehension question in application clothing — one that requires a student to identify which example from the lesson fits the scenario, rather than genuinely transfer the concept to a new one.
In practice, most exit quizzes need Recall → Understanding → Application. The Analysis level is useful for lessons at the end of a topic sequence — when students have enough content to evaluate and compare. Use it rarely and intentionally: it produces rich data but takes longer for students to complete and longer to scan.
Three checks between AI output
and tomorrow's lesson.
AI frequently generates Q3 application scenarios that resemble examples from the lesson — either directly or through obvious substitution ('same structure, different numbers'). Read Q3 and ask: could a student who memorised the lesson's examples answer this without understanding the underlying principle? If yes, replace the scenario with one that requires genuine transfer.
The diagnostic key is only useful if the named misconception matches what your class is likely to think. If your class has a specific prior misconception from an earlier lesson, replace the AI's generic wrong answer with the specific one. Generic diagnostic keys produce generic interventions.
AI sometimes answers Q2 ('explain why X causes Y') with a restatement rather than an explanation ('X causes Y because X is a type of Y'). Read the correct answer for Q2 and ask: does it explain the mechanism, or does it describe what happens? If it describes, rewrite it to explain.
The third AI tool:
differentiated activities in 2 minutes.
A3 covers AI differentiation — generating foundation, core, and extension versions of any activity from a single prompt. Combined with A1 (lesson regeneration) and A2 (exit quiz generation), it completes the daily AI toolkit for the agile teacher: improve yesterday's lesson, generate today's data collection, differentiate today's activities. Three tools. Under 10 minutes total.