Learn

Check understanding with Socratic questions

Probe definitions, reasoning, evidence, boundaries, and transfer one question at a time without supplying the answer

6 min setupTested with:ChatGPTReviewed: 2026-08-28

Add your context

Your text stays in this browser. AILesson Prompts does not send it to a model or server.

Topic and sourceLearner stateDialogue rules

Your prompt

Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt

Check the learner's understanding through a Socratic dialogue grounded in the supplied source.

Concept, course definition, examples, and objective:
[topic]

Level, prior explanation, current answer/confidence, misconceptions, and accessibility:
[learner]

Question count, difficulty, hints, feedback timing, source limits, and stopping rule:
[method]

Ask one neutral question at a time and wait for the learner's answer. Begin by eliciting their model, then probe definition, assumptions, evidence, counterexample, boundary, comparison, and transfer as appropriate. Do not hide a lecture inside the question, praise correctness before evaluating it, or reveal the full answer prematurely. After each response, reflect the learner's reasoning, identify the precise supported or questionable step, and choose the next question adaptively. Use the agreed hint ladder only after struggle. At the end summarize what the learner demonstrated, misconceptions still evidenced, source-grounded corrections, and one independent check.

Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.

From input to outcome

A worked example

See how concrete context turns this recipe into a usable result

Actual input

Topic and source: Opportunity cost: value of the next best alternative forgone when a choice is made; may be non-monetary. Objective: distinguish it from sunk cost and apply it to short scenarios. Source example: cheaper train may cost more when two hours of lost paid work are included. Sunk cost: already incurred and unrecoverable, should not affect current marginal choice.
Learner state: First-year student says: 'Opportunity cost is all the money you lose when choosing something.' Confidence 7/10. Often counts every rejected option. No accessibility needs stated.
Dialogue rules: Up to 6 questions, medium difficulty. Ask one at a time. Hint ladder: ask for a concrete alternative, then restate 'next best,' then provide the train example only if needed. Correct after two unsuccessful probes or an explicit request. Stop early if the learner correctly defines, distinguishes, gives a counterexample, and transfers.

Example output

First turn only: “You described opportunity cost as all the money lost. Imagine you choose to study for two hours instead of either working a paid shift or going for a free walk. Which one of those alternatives should determine the opportunity cost of studying, and what value would you count?”

If the learner says “both, because both were lost,” reflect and probe: “You have counted every rejected option. If the paid shift was your next best alternative and the walk ranked below it, what would change if the walk had never been available?” Wait again.

If still stuck, use hint 2: “The course definition says next best alternative, singular. Which alternative was next best?” Do not yet give a complete explanation. If corrected, probe distinction: “The shift payment has not happened. How is that different from a non-refundable ticket already paid?” Then ask for a non-monetary counterexample and a new transfer scenario.

End summary only after dialogue: record whether the learner identified one next-best alternative, recognized non-monetary value, separated forgone current alternative from unrecoverable past cost, and transferred. Do not treat confidence 7/10 as evidence of mastery.

Why this works

1
One-question turns reveal the learner's model before explanations can overwrite it.
2
Adaptive probes distinguish memorized definitions from boundary and transfer understanding.

Check the result

Does the dialogue wait for the learner after each single question?
Are probes neutral and responsive to evidenced reasoning rather than a fixed quiz script?
Are corrections delayed appropriately and grounded only in the supplied source?

More ways to explore