Read: ~9 min · Try it: 20 min
Not every assessment asks feedback to do the same job.
Sometimes feedback is meant to help students revise. Sometimes it is meant to justify a grade. Sometimes it is meant to diagnose what a class needs next. Sometimes it is part of a formal summative judgment where the stakes are higher and the room for adjustment is smaller.
The same feedback workflow should not be used in the same way for all of those situations.
This is where the distinction between formative and summative assessment matters.
Start with the purpose
Before using AI-assisted feedback on an assignment, ask:
What is the feedback supposed to do here?
If the purpose is to help students improve while the learning is still in progress, the assignment is functioning formatively.
If the purpose is to make a final judgment about achievement, the assignment is functioning summatively.
Many assignments sit somewhere between the two. A draft may be formative, while the final version is summative. A practice essay may prepare students for a formal exam. A graded assignment may still include feedback students use later.
The point is not to force every task into a category. The point is to decide what kind of support the feedback should provide.
Where TA39 is strongest
TA39 is strongest when feedback is:
- rubric-driven
- revision-oriented
- tied to clear criteria
- reviewed by the teacher before students see it
- used while students still have a reason to act on it
That makes formative and revision-based work especially strong use cases.
For example:
- a draft essay students will revise
- a practice response before a summative assessment
- a short analytical paragraph where students need targeted next steps
- a rubric-based writing task where feedback patterns repeat across a class
- a revision round where the teacher wants to see whether feedback was applied
In these cases, the value is not just speed. It is consistency, specificity, and the ability to make next steps visible while students can still use them.
Formative use: feedback as instruction
In formative work, feedback should point forward.
The most important question is:
What should the student do next?
That usually means:
- keeping next steps concrete
- limiting the number of priorities
- connecting comments to the rubric
- making feedback readable enough for the student to act on independently
- using Revision Rounds when students will submit another draft
A formative template should not try to say everything. It should help the student know where to focus next.
This is where feed-forward structures often work well: feedback organized around what the student should do next, rather than only what went wrong. They keep attention on revision.
Summative use: feedback as judgment and explanation
Summative feedback has a different job.
It may need to explain the grade, identify strengths and limitations, and connect the result to the rubric. Students may still learn from it, but they may not revise this piece of work.
That changes the review standard.
For summative use, be especially careful with:
- grade verification
- curriculum alignment
- rubric terminology
- consistency across students
- comments that imply a revision opportunity when there is none
If a score is being posted, the teacher needs to compare the descriptor, the student work, and the suggested grade carefully. A generated score is a starting point for judgment, not a verdict.
Summative work may still benefit from AI-assisted drafting, but the teacher review burden is higher.
When to use less
There are times when using less AI assistance is the better professional choice.
Consider using less when:
- the assignment is highly personal or sensitive
- the rubric is too broad to support fair drafting
- the task rewards originality or risk-taking in ways the rubric does not capture well
- the feedback would require extensive rewriting for most students
- the assessment is high-stakes and the review process cannot be rushed
- the teacher's knowledge of student context is central to the comment
Using less does not mean rejecting the tool. It means matching the level of assistance to the purpose of the assessment.
Sometimes the best use is full drafting and review. Sometimes it is using a template for structure but writing key comments yourself. Sometimes it is using the rubric tools before the assignment, but not using AI-generated feedback on the final product.
That judgment belongs to the teacher.
A practical decision framework
Before using TA39 on an assignment, answer five questions.
1. Is the rubric clear enough?
If the rubric is vague, the feedback will usually be vague too. Optimize or revise the rubric before running the assignment.
2. Will students act on the feedback?
If students will revise, resubmit, or apply the feedback soon, the use case is stronger.
3. How high are the stakes?
The higher the stakes, the more careful the review needs to be. For major summative work, plan more review time.
4. How much student context matters?
If personal history, confidence, language development, or classroom context matters deeply, plan for more teacher adaptation.
5. What will you do with the time saved?
If time saved goes into better review, student conferences, revision cycles, or clearer planning, the workflow is serving learning. If it simply compresses the process without improving quality, reconsider the setup.
Where TA39 fits
Use TA39 most confidently when the assessment is clear, rubric-driven, and connected to feedback students can act on.
Use it more cautiously when the task is high-stakes, highly personal, or difficult to capture in criteria.
The goal is not to use the same workflow everywhere. The goal is to make a professional decision about where AI assistance strengthens the work and where teacher-written feedback should lead.
Try it: classify your next four assignments
Estimated time: 20 minutes.
List four upcoming writing assignments. Use the five-question framework above as the structure for the exercise.
For each assignment, make one row with these columns:
- Feedback purpose — What is the feedback supposed to do?
- Revision opportunity — Will students act on it soon?
- Rubric readiness — Is the rubric clear enough to guide fair drafting?
- Stakes and context — How high are the stakes, and how much student context matters?
- Recommended use — full drafting, partial support, rubric/template prep only, or teacher-written feedback
Then choose one assignment where TA39 is likely to help most.
For that assignment, write:
- What should students do with the feedback?
- What kind of template will support that purpose?
- How much review time should be planned?
- Will Revision Rounds be useful?
- What would make this assignment a poor fit?
What to read or watch next
Watch Revision Rounds: Tracking Whether Feedback Lands if the assignment is formative or revision-based.
Read Hard Cases — Where Teacher Judgment Matters Most if the assignment is sensitive, complex, or difficult to capture in a rubric.
Read Assessing Your Own Growth with TA39 if you want to turn these decisions into a longer-term practice goal.
Closing thought
The best question is not, "Can I use AI feedback here?"
The better question is, "What is feedback supposed to do in this assessment, and what level of assistance helps me do that well?"
That question keeps the tool in the right place: useful, bounded, and guided by teacher judgment.