UI UX
Video Assessment
Designing evaluation around evidence, not just answers
Context
A video-assessment evaluation experience for faculty at partner colleges. Students record spoken responses, and faculty review and evaluate them against a rubric.
Faculty work through responses one at a time. An overview gives them a starting point, and a detail view brings together the student’s response, its supporting evidence, the AI’s evaluation and the rubric.
Who worked on it
The problem
Unlike a written exam, where answers can be skimmed, a video has to be watched in full.
The feedback we got was blunt.
“Faculty won’t scroll.”
Relayed by colleagues who spoke with faculty, not by faculty themselves.
It arrived as a rule rather than the behaviour behind it, and taking it at face value would have turned the design into a cramming exercise.
The real problem sat underneath it.
Faculty aren’t just watching a video. They’re comparing evidence against criteria and making a judgment.
Explorations and decisions
Keeping evidence and evaluation together
Our first version gave the response most of the space and hid the evaluation behind a disclosure. It looked clean, but grading meant constant movement between the two.
Putting the evaluation beside the response cost playback size, and that was the right trade.
Summary and Transcript
Alongside the video, each response carries a summary and a transcript, so faculty can go through it whichever way suits them.
AI-assisted scoring, with faculty in control
The AI pre-fills rubric scores rather than presenting a blank form, which turns scoring into editing rather than entry. Faculty adjust the overall score and the rubric scores rebalance.
Malpractice as evidence, not a verdict
The system surfaces what the AI observed, at specific timestamps, with a confidence percentage. It says this is what I saw and this is how confident I am, not this student committed malpractice.
The final design
Summary and transcript
The video sits first because it is the original evidence. The summary and the transcript are derived from it, so they stack underneath rather than beside it, and one control switches between them.
Every passage in the transcript carries its own timestamp.
The evaluation panel
The panel sits beside the response. Scores arrive pre-filled, and faculty adjust the overall score while the rubric scores rebalance.
Overriding the overall score switches the AI indicator off, so the mark reads as theirs. It is confirmed separately, and until then the AI’s score can be restored.
A short note justifies each AI score so faculty can see the reasoning behind it, and they can add their own written feedback alongside.
Malpractice observations
Each observation is pinned to a timestamp in the response and carries a confidence percentage, so it can be checked against the video rather than taken on trust.
The markers sit on the playback bar, so faculty jump straight to the moment rather than scrubbing for it.
One reads possible malpractice, 70% confidence, at 00:40, sustained off-screen eye movement.
What I measured
Timed in-house with and without the AI pre-fill: evaluation time per response fell by 60%.
Faculty said it felt quicker too. The older objective and subjective modules had no AI suggestion, so they read everything and typed every score themselves.
Still open: how often faculty open the malpractice timestamps, which says whether the flagging is used or skipped.