AI grading

AI-assisted grading that respects the teacher’s judgement.

AI proposes a first-pass review against the answer key and rubric. The teacher approves or overrides every mark. Nothing auto-publishes.

The teacher remains the author of record

AI is a fast first reader. It runs through each answer, matches it against the rubric and proposes a mark with a short justification. Then it stops. The teacher reads the AI’s suggestion and decides. Approve, edit, or override — every mark is the teacher’s.

What it’s good at, what it’s not

Objective and short-answer questions are where AI grading saves the most time. For long-form subjective answers, AI surfaces a draft mark and the specific rubric points it matched — the teacher still reads the answer, but skips the “where do I begin” cost.

A defensible audit trail

For every assessment, Relearns stores both the AI suggestion and the teacher decision, with timestamps. If a parent asks how a mark was arrived at, the school has a clean answer: AI proposed x, the teacher confirmed y, this is why.

Cost is visible, not hidden

AI cost is rolled up and surfaced per generate. Schools see the cost of every AI run and the cost-per-generate trend at admin level. There is no black-box meter.

Where this fits in the workflow

Grading sits downstream of teaching and assessment generation — see AI teaching material for how the practice paper and assessment are produced, and For CBSE schools for how the marks flow into PT, IA and the HPC.

AI grading FAQ

AI grading, answered.

AI runs a first-pass review of a student submission against the answer key and rubric. It proposes a mark and a justification for each question. The teacher approves or overrides each one. Nothing is auto-published — the teacher decides.