Use answer evidence to separate concept, procedure, interpretation, and careless errors before planning remediation
8 min setupTested with:ChatGPTReviewed: 2026-08-28
1
Add your context
Your text stays in this browser. AILesson Prompts does not send it to a model or server.
2
Your prompt
Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt
Diagnose knowledge gaps only from the supplied course context and answer evidence.
Course, level, objectives, source rules, and upcoming use:
[context]
Questions, learner working, answers, correct answers or rubric, and feedback:
[attempts]
Study time, depth, allowed materials, and solution policy:
[constraints]
For each error, quote or precisely locate the step that supports your diagnosis. Classify the most likely cause as concept, prerequisite, procedure, interpretation, recall, calculation, or execution slip, but do not treat a single mistake as proof of a stable gap. Separate observed evidence, hypotheses, and unknowns. Group repeated errors into underlying skills, show dependencies, and assign confidence. Propose one short diagnostic question per uncertain hypothesis before remediation. Then prioritize the smallest set of explanations, worked examples, retrieval practice, and transfer questions that fits the constraints. Withhold full solutions if requested. End with a reassessment rule that can confirm whether each gap has closed.
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
Course and target
Grade 8 linear equations. Objectives: distribute, combine like terms, preserve equality, and verify by substitution. Use only the supplied teacher key. Quiz retake is in 6 days.
Four 25-minute sessions. Diagnose before teaching. Do not reveal answers to diagnostic questions until I respond.
Example output
Evidence-linked diagnosis
Distribution across addition — medium confidence gap. Q1 changes 3(x+2) to 3x+2, so the multiplier was applied to x but not 2. Q3 correctly changes 2(x-4) to 2x-8, so this may be sign- or form-dependent rather than a fully missing rule.
Moving terms and combining like terms — currently demonstrated. Q2 is correct and Q3 preserves equality through the isolation steps. One item each is insufficient for mastery.
Verification by substitution — unknown, not a proven gap. No check is shown, but absence of written work does not establish inability.
Diagnostic questions, answers withheld
A. Expand 4(y+3) and explain what the 4 multiplies.
B. Expand -2(y-5) and explain each sign.
C. Without resolving Q1, substitute 13/3 into both sides and report whether they match.
Priority plan: session 1 uses A and B to distinguish general distribution from sign handling, followed by two interleaved expansions. Session 2 practices equations where distribution is and is not needed. Session 3 adds substitution checks and error explanation. Session 4 uses three unseen mixed problems. Mark the gap closed only if at least three differently formed distribution items are correct with an explanation and the learner independently rejects an invalid result by substitution.
Why this works
1
Separating evidence from hypotheses prevents one wrong answer from becoming an unsupported label.
2
Dependency-aware remediation addresses root prerequisites before repeated surface practice.
Check the result
Does every diagnosis point to specific answer evidence and state uncertainty?
Are careless slips distinguished from repeatable conceptual or procedural gaps?
Can the reassessment rule demonstrate that the learner transfers the repaired skill?
ChatGPT