Using AI · Growth Library
Give AI a Rubric and Ask for Specific Critique
Turn vague feedback into a check against explicit criteria
Try this
Write three to six criteria that describe a good result. Give the model the draft and rubric, ask for evidence from the draft for each criticism, and request no rewrite on the first pass. Triage the feedback, then make the changes you agree with.
Check-in: Did the critique point to observable gaps in the draft, or mostly produce generic advice I could paste onto any text?
Evidence: Practical guidance; no evidence-like claim is detected in the current source record.
Review note: Practical feedback workflow. The user defines the rubric, inspects evidence behind the critique and accepts or rejects changes; no universal superiority claim is made.
“Make this better” invites the model to choose both the problem and the solution. A rubric keeps at least one of those jobs with you.
1. Define what good means
Choose a short set of criteria before asking for feedback. If the criteria conflict, rank them or state the trade-off.
2. Require evidence from the draft
Ask the model to quote or point to the section behind each criticism. Unsupported feedback should be easier to reject than persuasive commentary.
3. Triage before editing
Mark each suggestion accept, reject, or inspect. Check factual or domain-specific advice against reliable sources rather than assuming the reviewer is right.
4. Revise with intent
Make the changes yourself or ask for a bounded rewrite of one section. Compare the result against the same rubric instead of changing criteria midstream.
Questions to consider
Why ask for critique before a rewrite?
A critique keeps the problem visible. An immediate rewrite can hide whether the model improved the intended criterion or simply changed the text.
What makes a useful rubric?
Use observable criteria tied to purpose: accuracy, required points, clarity for the target reader, evidence, tone, length, structure, or another condition you can inspect.
Maintenance record
Review history
Current status: Reviewed. An editorial or evidence review is recorded and no later event changes that conclusion.
2026-09-15 · reviewed · Brali editorial agent Source: review-registry-using-ai-starters.json
A bounded feedback workflow that keeps criteria and acceptance with the user; no unsupported claim that AI critique is universally superior to human review.
Browse the review ledger · Challenge or update this hack
Implementation observations can trigger a review, but they cannot change an evidence conclusion by themselves. Commercial relationships do not control review status or outcomes.
Continue from here
Choose the next useful path.
Article versions
Brali keeps substantive article history visible. Later reviews may refresh wording, sources, examples, or boundaries without silently replacing the record.
- Version 2026-09-15: first recorded publication in the migrated Brali corpus.
- Evidence review 2026-09-15: reviewed by Brali editorial agent; evidence state is
practical.
Canonical source record · Evidence state: practical.