Research & discovery methods:
- Journey mapping
- User interviews
- Interactive prototype for focus group
- Interactive prototype for usability testing
- Live alpha release pilot study
- Site analytics
- Post-pilot user interviews
Early conversations started with the question: How might we build an auto-scoring capability for writing that learns to grade like the instructor using it?
Using design thinking and user-centric research, the team began sketching out a plan for an MVP product.
More investigations were taken around questions of how to balance the needs of the pilot users and the goals of the AI & machine learning research being conducted.
The balancing act continued - What information do users need in order to maintain transparency during the "training" of the machine-learning algorithm? How much complexity should users be exposed to?
A round of user testing followed - to see what kinds of infographics resonated with instructors, and what information did they need to know when:
In the final analysis, simplicity won out over more complex communications. Here are the final designs for the instructors' assignment summary & grading screen
