How to grade written answers with AI assistance
Use AI scoring as a first draft. Review the answer, suggested score, and remarks before students see the result.
Prepare the assessment criteria
Before publishing, check each open question’s expected answer and points. Clear criteria give you a consistent basis for reviewing both student work and AI suggestions.
Request scoring from the results workflow
After students submit, open the exam results and use the AI scoring action for open questions. Scoring runs as a queued job; wait for completion before reviewing the updated results.
Review each suggestion
Check whether the response answers the question, whether a different but valid explanation deserves credit, and whether the remarks support the assigned score. Adjust scores and remarks where necessary.
Use a concrete review example
Suppose a student explains a process correctly using different terminology from the expected answer. Check the reasoning before accepting a lower AI-suggested score. This is an illustrative example, not a measured claim about grading accuracy.
Release results deliberately
Open-question exams keep results hidden initially. Release them after review. Students use the same exam link and their score code to see their own result.
Related workflows and guides
Review written answers with AI assistance
Use AI-assisted scores and remarks for open answers, review the suggestions, and release results when grading is ready.
How to run an exam without student accounts
Share an exam link, help students keep their score codes, collect answers, and explain when results become available.
Ready to create your first exam?
Start with your teaching material, review the questions, and share one student link.
Create your first exam free