Project overview
This project introduced an integration feature allowing school administrators to connect their contracted professional captioning vendors directly to Canvas Studio. Once connected, educators could request accurate, human-made captions, and admins could handle the requests without leaving the platform.
Problem statement
The Challenge
Originally, the only type of captions available were automatically generated ones, which had an accuracy of 85-90%. Professional and technical terms were often incorrect, and even simpler phrases could be inaccurate. Educators had to thoroughly review captions or risk publishing inaccurate ones, which was time-consuming.
The Goal
The aim was to improve caption accuracy and reduce the time educators spent reviewing and correcting captions. Captions made by humans are more accurate, reducing the burden on teachers and ensuring that captions are readily available for students.
Provide admins with a clear, cost-aware request management interface.
Research and Insights
User groups
While we did not specify detailed user personas, our design focused on two clear key groups:
School admins
Manage the integration and approve requests.
Educators
Request captions for their videos.
Key problems
We conducted interviews with 8 school districts across the US, involving both admins and teachers to understand their needs and challenges.
Inaccurate technical terminology
Identified urgent need among science teachers whose content relied heavily on precise terminology.
Review bottleneck
Discovered that some educators skipped reviewing captions entirely due to time constraints, or sometimes enlisted students to help.
External request management
Administrators had to approve or deny caption requests and track costs on the third-party vendor’s platform, with no way to manage them directly in Studio.
Design process
Ideation
User needs were clear – the solution required three core components:
An integration interface for school admins to link their vendor accounts to Canvas Studio.
A new professional caption requesting feature for educators.
A dashboard for admins to approve or decline caption requests.
Prototyping
We aimed for clarity and simplicity, avoiding clutter and unnecessary complexity. Initial wireframes quickly met functional requirements, allowing us to move into Figma prototypes.
User feedback
We conducted usability testing with the most insightful admins from our previous interviews. Initially, the dashboard used two tables, but feedback indicated this was overwhelming and reminiscent of Excel, – “we see enough Excel in a day”.
We switched pending requests to a card-based layout for a more approachable, scannable interface, while keeping table format for historical reviewed requests.
Solution
1. Integration interface
The interface offers the following features:
Vendor integration
Allows admins to link their third-party captioning vendors to the Canvas Studio instance.
Role management
Enables the creation and management of roles. Permissions include the auto-approval of requests, or providing only some type of services.
2. Caption requesting interface
The already existing caption requesting interface got a new feature:
New request type
New request type integrated into each media’s menu. Educators can specify the language and the service type, that mostly affects turnaround time and cost.
3. Admin dashboard
The interface offers the following features:
Pending requests
A card-based layout displaying each new request with details such as video length and name, making it easier for admins to review and decide.
Reviewed requests
A table format for handled requests, providing a clear histoical tracking.
Accessibility support
Accessibility was considered as soon as possible:
Ensured full keyboard navigation for integration setup and review process, with a logical focus order to support seamless navigation.
Defined ARIA labeling to all interactive elements for screen reader compatibility.
Results and Impact
This update for Canvas Studio resulted in the following positive experiences in the affected schools:
Enhanced accuracy
The professional captions significantly improved accuracy (from 85-90% to 98%+), reducing the need for teachers to review and correct them.
Streamlined workflow
Teachers saved considerable time by not having to correct captions, freeing time for teaching and content creation.
Legal compliance
Feature ensured adherence to U.S. accessibility laws for educational video captions.
Next steps
To ensure continuous improvement, the team plans to gather more feedback to enhance the feature further. The aim is to explore the development of new functionalities, such as automated approval based on cost management details, and more detailed analytics for administrators.
Personal highlights
Validated design decisions through direct contact with the feature’s intended end users.
Improved usability by ditching overwhelming tables, based on feedback.
Designed for scale, accommodating multiple vendors, service levels, and cost controls.
.svg)
.png)