Results are rarely perfect on the first try. If you are not sure how to get better results that are closer to your grading expectations, here are some tweaks that you can make on TimelyGrader to bring AI grading suggestions and feedback more in line with your expectations.
Step 1. Always start with a baseline (even if it's not accurate)
The goal is to create a baseline and get your eyes on the first iteration. Accuracy isn't important at this stage.
Start by uploading your assignment and rubric with minimal edits to create a baseline. Import or upload assignments to run a first pass to see how the rubric is interpreted by TimelyGrader. We will show you grading explanations alongside suggested grades so you can spot issues like vague wording (“good analysis,” “clear argument”), criteria that don’t state intent, or missing context in separate documents.
Treat these as opportunities to improve and use the baseline to spot hallucination or misinterpretations. Finding errors is the first step.
Example:
An instructor uploads a “Marketing Strategy Report” with “Recommendation” as a criterion and four ratings (excellent, good, fair, and poor). Because “Good” isn’t defined, the system can’t tell what it means to be good, fair, or poor so it tries to take a guess, which often than not, deviates from expectations.
Step 2. Identify What Needs Adjustment
When grading results don’t match your expectations, you can improve accuracy in two main ways:
a) Update the Rubric
Sometimes the rubric is too general or hard to read the same way every time. Improve yours by adding clear performance descriptions for each rating level, replacing vague phrases (like “good recommendation”) with measurable indicators or adding criteria to match what you care about now. The more concrete the rubric, the more consistent the grading.
Example:
Before: “Excellent Recommendation — Presents a strong recommendation with relevant sources.”
After: “Excellent Recommendation — Presents 3 clear, evidence-based, and actionable recommendations that use relevant sources to support key claims.”
b) Add Specific Grading Instructions
If you have grading notes for special cases and judgments, add them to TimelyGrader as specific instructions so the AI follows your intent without revealing answers to students. Use clear and direct rules that resolve ambiguity (for example, when to give partial credit, how to handle formatting issues, or what counts as acceptable sources). This added context improves fairness and consistency by aligning AI behavior with your real grading logic.
Example:
Instruction: “If the student does not make this particular recommendation to the case study, do not provide full marks regardless of the recommendation”
Result: The AI will only grant full marks if the students make a particular recommendation to the company.
Step 3. Iterate and Regenerate
After making changes, you can reset submissions in TimelyGrader. Apply your updated rubric or instructions, and regenerate to measure improvement. Don’t be afraid to repeat in small cycles until results align with your grading standards!
Example:
After adding clearer descriptors and a few specific grading rules, the instructor regenerates. The AI now distinguishes “excellent recommendation” from “good recommendation,” and creating more distribution in the grading suggestions.
Excellent recommendation: “The student presents a strong and correct recommendation as per the grading instructions with more than 4 actionable steps for the company to take. The student also includes more than enough sources to back up potential benefits and risks.”
Good recommendation: “The student makes a compelling recommendation, although incorrect as per the grading instructions, but includes significant reasoning and sources to back up their recommendation."
Step 4. Only you know how you grade
AI grading refinement works best when paired with your subject matter expertise. Use the AI to detect patterns and inconsistencies, then decide whether the issue is vague rubric language or missing context. Each iteration moves you toward a grading approach that’s efficient, consistent, and aligned with your standards.
Example:
A writing instructor sees the AI being too lenient on grammar. They adjust rating descriptions to make it harder to get higher grades so the next pass matches what they would have given.
Bottom Line
TimelyGrader isn’t automation. Treat it as a teaching assistant that you need to train and that improves with your refinements. With each iteration and clearer instructions, TimelyGrader will reflects your academic judgment more accurately, saving time while maintaining quality and depth for your students.