Every Report Card Comment I Wrote Started With the Same Three Words

I noticed it on the printout, not while I was writing. Six comments in a row, all starting with some version of "continues to show." Continues to show strong effort. Continues to show growth in reading comprehension. Continues to show a positive attitude toward learning. Six different students, six sentences that could have belonged to any of them.

Report card comments have a specific failure mode that essay feedback does not, and it is worth naming clearly before getting into how AI helps or does not help with this particular task. An essay comment has room to be specific because you are usually writing eighty to a hundred words about one piece of work you just read. A report card comment is often one or two sentences, sometimes with a hard character limit set by your school's system, covering an entire term of a student's work across every subject or a specific subject over several months. There is very little room to hide genericness in two sentences, and there is very little room to be specific either, which is exactly the tension that makes this task harder than it looks and exactly why fatigue shows up faster here than in almost any other writing task teachers do.

Why AI Generated Report Card Comments Fail in a Specific Way

Ask any AI tool to generate a report card comment from a vague prompt describing a student's general performance level, and you get output that sounds like a report card comment without actually describing anything specific to that student. This happens for the same reason generic quiz distractors happen and generic rubric language happens. The model is producing the statistically likely version of the category you asked for, and report card comments as a category are dominated online by exactly the kind of safe, vague, universally applicable phrasing that makes for a bad individual comment.

The specific words to watch for, because they show up constantly in AI generated report card comments regardless of which tool you use: continues to show, is a pleasure to have in class, demonstrates a positive attitude, works well with peers. None of these are wrong. All of them could describe almost any student in almost any class, which means none of them are actually doing the job a report card comment is supposed to do, which is telling a specific family something true and useful about their specific child.

The Input That Actually Fixes This

The fix is not a better tool. It is being specific about what you feed the tool before it writes anything, the same principle that runs through every other writing task on this site, but the stakes for skipping it are higher here because you have so little room in the final output to correct for a vague start.

Before generating any comment, write three things about the specific student in your own plain language. One specific thing they did this term that you actually remember, not a general trait. One specific area where they are still developing, described concretely enough that a parent would know exactly what it means. And one specific thing you want the family to try or notice at home, if that fits your school's comment format.

Here is what that looks like as an actual prompt, using an invented but realistic example.

Write a report card comment of two sentences for a student. Specific detail one, the student led her small group through a disagreement about how to organize their science project data without needing me to step in, which is new for her this term. Specific detail two, her written explanations are accurate but often skip the step of explaining why a result matters, jumping straight from data to conclusion. Detail three, at home it would help to ask her to explain her reasoning out loud before writing it down. Write this warmly but specifically, in a tone appropriate for a parent audience, and do not use the phrases continues to show, is a pleasure to have in class, or demonstrates a positive attitude.

That explicit instruction to avoid the specific generic phrases matters as much as the specific details themselves. Naming what to avoid is often what stops a tool from defaulting back to safe language even when you have given it something specific to work with.

What Good Output Actually Looks Like

Using the prompt above, a genuinely improved comment reads something like this. She showed real leadership this term, navigating a group disagreement about project organization without needing adult support. Her writing is accurate but tends to state conclusions without explaining the reasoning behind them, so practicing explaining her thinking out loud before she writes would help her writing catch up to her thinking.

That comment could not be swapped onto a different student's report card without becoming obviously wrong, which is the actual test of whether a report card comment is doing its job. A parent reading it learns something true and specific about their child that they did not already know from a grade alone.

Where This Still Requires Real Care

Even with a specific prompt, review every generated comment for two things before it goes into a report card, since the character limits and formality of this format create specific risks that longer feedback formats do not.

First, check tone very carefully around anything describing an area of growth. In a short comment with little room for softening context, a developing area can read more harshly than intended if the AI's phrasing is even slightly more blunt than you meant. A sentence that would read as gentle and specific at eighty words can read as clipped and critical at fifteen.

Second, verify the comment does not accidentally use language associated with a diagnosis or clinical label the student does not actually have. A comment describing a student as having difficulty focusing, when generated from a vague prompt, occasionally drifts toward clinical sounding phrasing that implies more than a teacher's classroom observation actually supports. Keep the language grounded in what you specifically observed, not a broader characterization of the student as a type of learner.

The Actual Time Difference

For a class of twenty five to thirty students, writing genuinely specific comments from scratch typically takes a full evening when done carefully, longer if done at the end of a marking period alongside everything else competing for that same time. Using this workflow, meaning writing three specific details per student first and then generating from those details, the writing itself takes closer to five minutes per student rather than ten to fifteen, because the thinking work of what to say is separated from the writing work of how to phrase it well.

The thinking work does not get faster. Knowing what to say about a specific student still requires knowing that student, which is not something any tool changes. What gets faster is turning that knowledge into well phrased, appropriately toned language across an entire class without the sixth comment sounding like the first one, tired and slightly less specific than it should be.