Open each draft, check the scores and comments against your own rubric, then edit anything that misses the mark before you approve it. JeddAI drafts feedback; you stay the reviewer. Nothing reaches a student until you have read, adjusted, and released it yourself.
What does reviewing JeddAI's draft feedback actually involve?
Reviewing means reading each drafted score and comment, comparing it to your rubric, and editing anything that misses before you approve it. JeddAI produces a first draft for every piece of student work, and your job is to check it rather than accept it. The draft handles the routine writing, so your attention goes to judgment instead of transcription.
In practice the workflow is a single pass per submission. You open a draft, confirm the grade reflects the criteria you set, read the comments for accuracy and tone, adjust what needs adjusting, and release it. Across a class set most drafts need only light edits and a few need real rework, which is exactly where your professional judgment earns its place.
What should I check first when reviewing a draft?
Check accuracy before anything else: does the score match your rubric, and do the comments point to real evidence in the student’s work? A polished comment attached to the wrong score helps no one, so confirm the judgment first and the wording second. Read the drafted grade against your criteria, then read the comment that is meant to explain it.
Work criterion by criterion where you can. If your rubric weights argument and evidence heavily, make sure the draft graded those the way you would, not just the surface features that are easy to spot. Where a comment references a quotation or a paragraph, glance at the student’s work to confirm the reference is real before you approve it.
- Score: does the grade match the performance level your rubric describes?
- Evidence: do the comments cite something that is actually in the student's work?
- Coverage: are the criteria you care about most addressed, not just the easy ones?
- Tone: is the feedback specific and usable, rather than generic praise or vague criticism?
- Actionability: does each student get at least one concrete next step?
How do I edit the wording so the feedback sounds like me?
Edit the wording by rewriting anything that does not sound like how you talk to your students. The draft applies your criteria, but the voice stays generic until you shape it. Change phrasing that is too stiff, too soft, or too long, and cut praise that does not point to something specific. Feedback students act on tends to be short, direct, and tied to a next step.
Comment banks are the fastest way to keep that voice consistent. If you have saved the phrases you reuse for common strengths and errors, the draft pulls from them and you edit less. Where a saved phrase almost fits, tweak it in place; where you keep rewriting the same idea, add it to the bank so the next draft starts closer to your own wording.
Keep the student in mind, not just the rubric. A comment that names a criterion is defensible, but a comment that tells a Grade 9 writer exactly what to do next is what moves their work forward. Editing for that second quality is where your time pays off most, and it is work only a teacher who knows the class can do well.
How do I catch mistakes before students see them?
Catch mistakes by reading the draft the way a skeptical colleague would, looking for claims the student’s work does not support. Like any AI, JeddAI can misread a passage, credit an argument that is not there, or miss one that is. Because you review every draft, these slips are caught at your desk rather than in a student’s hands.
Pay closest attention where the stakes are highest: the overall grade, any comment that makes a factual claim about the work, and feedback for a student who may be surprised by it. If a score feels off, re-read the student’s response before you trust the draft’s version of it. The teacher-in-control step exists precisely so no unchecked judgment reaches a learner.
How much time should I spend reviewing each draft?
Match your review depth to what the assessment is for. A low-stakes practice piece might need a quick scan to confirm nothing is wrong, while a summative grade that lands on a report card deserves a close read of every comment. The rubric holds the baseline steady either way, so you can lean harder on the draft when the stakes are low.
This is why the first pass saves time without lowering standards. On routine work you accept most of the draft after a glance; on high-stakes work you read closely and rewrite freely. Either way you are editing a draft rather than composing from a blank page, and the time you save on routine feedback is time you can give to the students who need a real conversation.
Do students ever see feedback I have not approved?
No. Students only see feedback you have reviewed and released, and nothing is published automatically. JeddAI drafts scores and comments into a space only you can see, and each draft stays there until you approve it. You can edit any part, change the score, discard a draft entirely, or write your own comment in its place, and the draft never moves to a student on its own.
This is the core of the teacher-in-control design. The tool does the drafting; you remain the grader of record, and the feedback a student reads carries your judgment, not an algorithm’s. That distinction matters for trust: when a student, parent, or administrator asks how a grade was reached, the answer is that you reviewed and approved it against your rubric.
How do I make reviewing faster over time?
Reviewing gets faster when you feed what you learn back into your setup. Every time you rewrite the same comment, notice which criterion or descriptor was unclear, or reach for a phrase that is not in your bank, you have found something to fix. Sharpen the rubric, add the phrase, and the next batch of drafts arrives closer to what you would have written.
A tool like JeddAI helps you apply the same criteria evenly across a class and cut the time spent drafting routine feedback, while you review and approve every score and comment. Built in Australia and expanding internationally, it keeps the teacher in control. You can Get started with JeddAI and try the review workflow on a single assignment before scaling it up.
| Aspect | Low-stakes practice work | Summative or reported grade |
|---|---|---|
| How closely to read | Quick scan for obvious errors | Read every score and comment closely |
| What to prioritize | That nothing is plainly wrong | Accuracy, tone, and fairness on each piece |
| Editing effort | Accept most, fix the odd comment | Rewrite freely so the feedback lands |
| Main risk if you skim | A minor unhelpful comment | An unfair grade a student may contest |
Frequently asked questions
Can I edit the score itself, or only the comments?
Both. You can change any score as well as the comments, so if the drafted grade does not match your rubric you override it before approving.
Can I approve feedback for a whole class at once?
You review each piece before it is released, since students differ and drafts vary in how much editing they need. Reviewing individually is what keeps the feedback fair and accurate.
Does editing a draft train JeddAI to write more like me?
Not automatically. You shape future drafts by updating your rubric, success criteria, and comment banks, which is what the next draft draws on, rather than it learning silently from your edits.
Is there a limit to how much of a draft I can change?
No. You can adjust a single word or rewrite everything, the score, the comments, or both. The draft never constrains your final judgment.
Can students tell their feedback began as an AI draft?
The feedback students receive is what you reviewed, edited, and approved, so it reflects your judgment and voice rather than a raw draft.
Get started with Jeddle
Jeddle gives teachers and students instant, syllabus-aligned feedback powered by JeddAI.
Looking for study material? Browse Jeddle's Australian-English subject resources, or explore more articles on Using Jeddle.
