Class Companion is an AI EdTech platform that enables students to receive instant AI feedback on their written assignments. Through Class Companion, teachers add assignments, and students complete those assignments.
Class Companion was preparing for Series A funding with a critical need to improve key metrics. Our data revealed a significant opportunity: 18% of assignments were incomplete and 3% not started at all, directly impacting our ability to demonstrate platform effectiveness to investors.
After reviewing FullStory and Mode, I reviewed the existing student submission page and identified general design improvements.
Typography updates: Source Sans was selected to replace Inter font after testing for better x-height, multiple weights, and italic support—significantly enhancing readability for students working with dense content and complex data tables.
Rapid iteration: We explored connected and disconnected chat boxes, scores in the left and in the right columns, assessment criteria opened and closed, and more.
Improvements: These relatively small design changes increased readability and understandability.
Improved navigation: A sticky header shrinks on scroll, maintaining context while maximizing space.
Disputed scores: Students need to be able to dispute scores given by AI. Because this functionality would eventually be unnecessary due to improved scoring algorithms, we kept the interaction areas small.
Error handling: Besides standard error states on the page, we also wanted to let students know when they had remaining attempts on a question with a non-perfect score.
Header context: The submission page needed to work for both teachers and students. One clear improvement from the previous implementation was updating the header for teachers — before, the H1 was the assignment title. However, teachers view assignments based on the submission from a specific student, so we updated that H1 to the student’s name
