Platform guide

Grading & oversight

Human-in-the-loop grading, handwritten work with OCR, and communicating grades with students.

Section videos

Walkthrough videos for Grading & oversight

Pick a walkthrough to play it. Worth watching before working through the guides below.

An open-ended question with its Grading panel expanded. Each criterion, such as “Identifies effects on internal users”, carries a percentage weight, and criteria can be generated, added, locked or removed.

Written guides

Collage AI provides immediate automated scoring suggestions for up to 98% of submissions, but the engine runs entirely under a human-in-the-loop framework: instructors hold the final editorial mandate over points and notes.

The core statistical spot-check. Target audits at mid-range scores (e.g. answers receiving 70–80%) or multi-point open prompts. High- and low-scoring tiers usually yield clear-cut matches.
Criteria group sorting. Sort analytics trends by specific criteria lines rather than student by student. This quickly exposes whether a foundational question needs rubric refinement.
Immediate point rewriting. Clicking directly into any active point or evaluation field instantly replaces the automated baseline, adjusting individual performance sheets in real time.

Educator FAQ