Media staff
Needed a faster way to organize large libraries, find issues, and avoid rebuilding the same review setup for every video.
03Human-Computer Interaction hackathon 2025 · Team InSight
Video Accessibility Assistant
A human-reviewed workflow for helping university teams find, correct, and track video accessibility issues at scale.
- Role
- UX design team member
- Team
- Team InSight
- Methods
- Needs analysis, workflow mapping, interface design
- Deliverable
- High-fidelity prototype and pilot recommendations
The problem
A growing accessibility backlog made a fully manual process difficult to sustain.
One Iowa State department had more than 1,000 videos requiring scene-level descriptions before a December compliance deadline. Staff had to inspect each meaningful scene change, write descriptions, and review the full timeline.
Commercial tools were too expensive to scale campus-wide, while different departmental processes created duplicated effort and inconsistent quality. The design challenge was to reduce repetitive work without removing human judgment.
Who we designed for
The workflow had to support both production staff and the students relying on the result.
Faculty
Needed clear ownership, deadlines, and a way to understand whether course material was ready to publish.
Accessibility officers
Needed consistent checks and an auditable review process across departments.
Students
Needed descriptions and captions that accurately communicate essential visual information.
The response
Automate the first pass, then make every final decision reviewable by a person.
Detect and organize
The concept detects scene changes, analyzes captions and audio, drafts descriptions, and groups issues into a single review queue.
Review and approve
Staff see the flagged moment in context, compare the suggestion, revise it when needed, and confirm readiness before export.
Prototype evidence
Four screens show the complete workflow from project setup to accessible export.
The gallery follows the working sequence: track the library, analyze a video, review individual issues, and confirm export readiness.
Artifact 01
Campus project dashboard
A shared workspace helps departments track video volume, deadlines, ownership, and accessibility progress.
Ownership and deadlines stay visible across teamsLibrary-level progress helps staff prioritize work
01 / 04
Proposed pilot measures
We defined targets to evaluate speed, quality, cost, and adoption.
Target processing time per minute of source video.
Target description accuracy compared with human-reviewed quality.
Target cost reduction compared with commercial tools.
A shared process designed to support multiple departments.