Product Thinking
Canvas Course Authoring Workbench
A better authoring environment emerged from firsthand friction with live Canvas editing and evolved into organizational tooling.
EdPlus at ASU
- FastAPI
- React
- TipTap
- Claude
Measured
4 designers
Interviewed about their real authoring workflow before any interface decisions were made.
Measured
Review first
Nothing published to Canvas without an explicit approval step.
The Friction
Building and editing directly in Canvas is slow, and there is no safe place to work before publishing. Designers were making changes in the live environment because there was nowhere else to make them.
What I Did First
I interviewed four instructional designers about how they actually build in Canvas, not how the process is documented. Their answers drove the interface: the editing controls, the drag-and-drop reorganization, and above all the review step before anything reaches the live course.
What I Built
A separate build-and-review environment with a rich text editor, slash commands, drag-and-drop reorganization of pages and modules, and an AI side panel that generates a module from a description. Edits made in Canvas synced back automatically with no manual resync. The accessibility checks from the linter extension were built in. Nothing reached Canvas until it was reviewed and approved.
What Came of It
Built before the AI taskforce existed. Editing features from it, including the rich text editor and slash commands, were carried into another high impact project. Its successor, ID Workbench, added alignment checks, QA audit, and needs analysis grounded in learning theory and performance improvement frameworks, and was merged into the AI Course Authoring System.
Scope and Limits
This was a prototype, not a supported product. Its value was in proving the workflow and informing the tooling that followed.
More systems, and the learning experiences behind them.
or write directly to brent.michael670@gmail.com