Learn with Chat
Timeline
14 months
Size
5 people
Platform
Web (responsive)
Timeline
14 months
Size
5 people
Platform
Web (responsive)
Most Ed-Tech apps deliver generic content without verifying understanding, leaving students with a undifferentiated experience that isn't grounded in their unique background.
Learn with Chat turns a student's own materials into a chat-based tutor that gives personalized explanations, knowledge checks, and low-friction learning flows.
As the founding Product Designer, I owned design and implementation end-to-end, from concept to live prototype. I combined product design, front-end development, and AI-assisted workflows to rapidly iterate on learning interactions, personalization logic, and conversational UX.
Early wire-frames explored Learn with Chat as a structured learning environment, using calendars, progress indicators, and course modules to externalize study planning and accountability. As these concepts were tested through discovery questionnaires (n=32) and moderated usability sessions (n=8), a clear pattern emerged for how learners actually engaged with the system. Rather than navigating dashboards or managing modules upfront, participants treated the interface as a conversational entry point: initiating learning through questions, clarifications, and follow-up prompts.
Planning and progress views were interpreted as reference layers, while chat became the primary locus of sense-making. Over time, learners sought structure from the conversation, asking the AI to generate quizzes, summarize topics, or suggest next steps based on prior exchanges. This flow reframed structure as an emergent property of interaction. The resulting design centers chat as the generative surface of the product, allowing organization, assessment, and progress tracking to emerge contextually rather than as pre-configured workflows
After two rounds of A/B testing on 8 participants, aged 18 to 24.
1. Decompose courses into shared modules
I added modules as a partner concept to reduce duplication in learning and better pinpoint weaknesses. The course creation flow automatically analyzes materials and auto-creates modules to preserve flexibility while not adding to onboarding friction.
2. Immediate feedback over layered gamification
I removed gamified elements that required additional steps to view progress or grades, such as separate result pages and animations. While motivating for some users, these elements added latency and were frequently skipped during repeat use. Surfacing feedback inline enabled quicker comprehension and smoother continuation.
3. Interactive course maps over static outlines
I invested in an interactive course map to encourage exploration and engagement, despite higher implementation cost. A/B testing showed increased interaction compared to static outlines, leading us to allocate additional budget and engineering effort to this experience.