Output Rubric Designer
Use this to define measurable criteria for evaluating an AI-generated output.
Use case: A trainer, team, or individual can judge drafts consistently and compare prompt versions using the same standard.
Best tools
- ChatGPT — Builds task-specific rubrics and applies them iteratively to sample outputs.
- Claude — Provides nuanced criterion definitions and evidence-based evaluations of long outputs.
Fill in your details
Answer the questions below. Your answers are placed into the prompt automatically.
Your filled prompt (preview)
You are my Output Rubric Designer. Define measurable criteria for evaluating an AI-generated deliverable.
========================================
PERSONAL CONTEXT (filled by me)
========================================
Deliverable to evaluate: [DELIVERABLE]
Audience judging quality: [AUDIENCE]
Must-pass criteria: [MUST_PASS]
Nice-to-have criteria: [NICE_TO_HAVE]
Scoring scale to use: [SCALE]
Examples of failure: [FAILURE_EXAMPLES]
Weighting preferences: [WEIGHTS_HINT]
How the rubric will be used: [USE_CASE]
If any field is blank or still shows brackets, list the gap, continue with labelled assumptions, and never invent unverified facts, numbers, citations, or quotes.
========================================
WORKFLOW YOU MUST FOLLOW
========================================
STEP 1 — Criteria
Turn [MUST_PASS] and [NICE_TO_HAVE] into observable rubric rows for [DELIVERABLE].
STEP 2 — Scale
Define [SCALE] descriptors with concrete anchors, including [FAILURE_EXAMPLES].
STEP 3 — Weights
Propose weights from [WEIGHTS_HINT] that total 100.
STEP 4 — Usage
Explain how to apply the rubric in [USE_CASE] for [AUDIENCE], including AI-assisted pre-score caveats.
========================================
OUTPUT FORMAT
========================================
Use these exact headings:
A. Rubric table
B. Scale descriptors
C. Weights
D. Fail-fast gates
E. Worked scoring example
F. Common scoring mistakes
G. AI pre-score caution
H. Revision priorities
Rules: measurable over vague; no personality scoring; keep criteria observable.
You are my Output Rubric Designer. Define measurable criteria for evaluating an AI-generated deliverable. ======================================== PERSONAL CONTEXT (filled by me) ======================================== Deliverable to evaluate: [DELIVERABLE] Audience judging quality: [AUDIENCE] Must-pass criteria: [MUST_PASS] Nice-to-have criteria: [NICE_TO_HAVE] Scoring scale to use: [SCALE] Examples of failure: [FAILURE_EXAMPLES] Weighting preferences: [WEIGHTS_HINT] How the rubric will be used: [USE_CASE] If any field is blank or still shows brackets, list the gap, continue with labelled assumptions, and never invent unverified facts, numbers, citations, or quotes. ======================================== WORKFLOW YOU MUST FOLLOW ======================================== STEP 1 — Criteria Turn [MUST_PASS] and [NICE_TO_HAVE] into observable rubric rows for [DELIVERABLE]. STEP 2 — Scale Define [SCALE] descriptors with concrete anchors, including [FAILURE_EXAMPLES]. STEP 3 — Weights Propose weights from [WEIGHTS_HINT] that total 100. STEP 4 — Usage Explain how to apply the rubric in [USE_CASE] for [AUDIENCE], including AI-assisted pre-score caveats. ======================================== OUTPUT FORMAT ======================================== Use these exact headings: A. Rubric table B. Scale descriptors C. Weights D. Fail-fast gates E. Worked scoring example F. Common scoring mistakes G. AI pre-score caution H. Revision priorities Rules: measurable over vague; no personality scoring; keep criteria observable.
Pro tip: Weight factuality and safety separately from style so polished errors cannot score highly.
Expected output: A weighted rubric with anchors, failure gates, and a calibration exercise.