UI/UX, Interactivity, Web Design
CorpusKey
Improving an AI EdTech platform where professors can upload sources such as PDFs, chapters, or website links and CorpusKey will create a 'textbook' using all the information the professors have inputted.
My Role
UX/UI Designer
Duration
Aug. 2024 - Dec. 2024
Documentation
Presentation
Overview
CorpusKey is a web-based AI tool that helps instructors generate textbooks, presentations, and case studies built through the collection of files that an instructor feeds to the system. The tool also aims to quickly and efficiently grade language-based content without worrying about subjectivity and generate course materials like readings, tests, and presentations using natural language processing (NLP) capabilities.
The main focus of this project was developing user-flows for an improved experience for instructors, thus creating visual and interaction design demonstrating the grading system and onboarding.
Our team consisted of 8 people thus we split into two teams. I was part of Team A where we created a journey map and improved the interface, flow, and interactivity of several screens. Additionally, we designed an onboarding process that would make it easier for teachers to start using the website. Our teams collaborated often to give each other feedback and insights, and to ensure our designs were visually cohesive.
Research
Team A's original aim was to work on adding a file repository to the CorpusKey website. However, while conducting our initial activities, we realized that there was a greater issue in the overall navigation and flow. Thus, we pivoted to focusing on both the content generation process, the repository, and the onboarding.
From the findings of our activities, we decided to focus on the issues pictured below.
Design
We followed an iterative design cycle: develop wireframes, test, design further, test again.
My team focused on the creation of three screens. A repository screen, a content outline page, and the results of the AI-generated content page.
We whiteboarded and sketched out all our ideas first, utilizing insights from our research. Then we narrowed down our focus and moved towards creating mid-fidelity screens that we used for testing with users (instructors). Concurrently, we ideated ways to conduct onboarding (via pop-ups or notifications or a tutorial).
Results
My team's final results were successfully creating onboarding pop-ups, redesigning screens, adding interactivity, and ensuring our designs were consistent with Team B's work.
Repository Screen
The user (professor) will upload their files onto the repository screen. Documents and images are stored separately.
OUTLINE IMage
Now, teachers and professors can attach specific files or images to their overall content outline. The AI will use the attached document to create content for that particular part of the outline.
onboarding pop-ups
We added icons that when pressed would be a pop-up outlining the function of each element on the screen. This serves to guide the user and lessen confusion at each step of the phase. It also advises the user what to do to continue creating their content.
System Status
Users can now see how much progress is left until their content is generated. They can also cancel the generation anytime.
