Integrating AI in Canvas Quiz Creation
Redesigned the Canvas quiz creation flow, introducing autosave, AI quiz generation, and progress indicators.
Research and prototyping led to 62% faster task completion and 30% fewer clicks.
Identifying the Gaps
We interviewed 8+ educators and faculty to identify key issues. We then validated and explored these issues through observation interviews, measuring: time on task, click count, and user satisfaction.
- Worrying regarding saved progress
- Constant Repetition
- Mentally demanding
- Time on Task
- Click Count
- Error Rate
- Task Completion Rate
Problems Identified
Creating quizzes in Canvas for faculty is inefficient, confusing, and requires excessive time due to:
Lack of progress visibility
Redundant manual actions
Poor feedback mechanisms
User Journey while using Quiz creation on Canvas
"The system hasn’t been updated in years. There is the general feeling that the world has passed it by."
- Professor, Teaching Faculty
Our Challenge
Streamline the quiz creation process by:
Reducing time spent doing repetitive tasks such as; manually entering similar question types, adding settings for each question, saving repeatedly and,
Enhancing the overall experience for educators using Canvas.
Brainstorming Concepts
We conducted Low, Mid and High fidelity testing with faculty and measured the same metrics across testing to evaluate impact. These sessions provided tangible insights, allowing us to iterate and refine the experience based on faculty behavior and feedback.
How We Solved It
Through our iterative process, we developed solutions in the user flow and interface to address the pain points faced by professors.
Following are the 3 solutions:
1) Integrating AI in the Quiz Creation Process
How can AI speed up Quiz Creation and Repetition?
Through primary and secondary research, we evaluated the impact of AI on quiz creation. Studies showed educators completed quizzes in under 10 minutes and reported higher confidence using AI. Our faculty surveys highlighted strong interest in automating repetitive tasks.
Feedback from low and mid fidelity testing confirmed that “Build with AI” directly addressed key pain points like cognitive load, time constraints, and creative fatigue. Here are some Faculty quotes:
2) Including the Progress Bar
3) Reduce Steps and Time
Testing
We compared our baseline metrics to our metrics collected after high fidelity testing. The redesign led to a 62% reduction in time on task and 30% fewer clicks during quiz setup.
Key Takeaways
This project highlighted the power of thoughtful, user-centered design. By testing across fidelity levels and measuring performance consistently, we were able to make clear, data-backed improvements, reducing time on task by 62% and simplifying interactions by 30%.
We also learned that AI can be genuinely useful when it’s introduced with care. Faculty responded positively to assistive features like “Build with AI” because they felt optional, editable, and intuitive, not intrusive.
Most importantly, we saw how small changes, autosave, visual hierarchy, and progress indicators, can dramatically reduce cognitive load. Designing for clarity, control, and trust led to a tool that not only worked better, but felt better to use.
