VideoTutor · Phase 3
I led VideoTutor from a set of promising AI capabilities to a coherent learning product. I defined the end-to-end journey, initiated a strategic redesign, and built interaction patterns that could support new audiences and learning tools without fragmenting the core lesson.
Role
Lead Product & UX/UI Designer
TIMELINE
Apr - May 2026
scope
Product Strategy · AI Interaction · Shipped UX/UI
Impact
A scalable foundation for the evolving learning product
00
PROJECT context
The product opportunity
Generate to Animation
Long-term Memory
MY CONTRIBUTIONS
0–1
Establish the core
Define one complete journey from first question to continued learning.
1
Proactive redesign
Improve the learning loop
Evolve the experience for a broader audience and future product direction.
2
System extensions
Add capabilities gracefully.
Integrate downloads and exam tools without fragmenting the lesson.
3
01
Establishing the core
I led the complete experience across the landing page, logged-in return page,
lesson generation, interactive classroom, and future learning continuity.
BEFORE
A one-time learning journey.
Users typically arrived with a question, watched an explanation, and left. Once the lesson ended, there was little reason to return.
AFTER
An ongoing learning relationship.
Instead of relying on a single action to start learning, I redesigned the return experience around multiple entry points, visible progress, and emotional motivators to encourage learners to continue and come back.
Previous Iterations
V0 - MVP (before I join)
Validating the core loop
At the MVP stage, the product experience focused on the essential loop: ask a question, generate a video, and watch the explanation.
V1 - Early Beta Launch
Exploring reasons to come back
As VideoTutor entered early beta, we began experimenting with lightweight retention features, including recommended videos and saved content, to encourage users to revisit and continue learning.
V2 - Product Evolution
Simplifying the return experience
As features expanded, the experience was reorganized into dedicated spaces for learning, personal history, and community content, maintaining clarity while making room for future growth.
I designed the return page around four retention hypotheses:
01
Discover
Suggest topic-aware prompts and related facts.
02
Practice
Create a lightweight daily challenge habit.
03
Progress
Make streaks and tutor connection visible
04
Continue
Help learners resume recent lessons quickly.
Design judgment: I treated these as retention hypotheses, not proven gains. Their value at this stage was helping the team move from a one-time answer model toward a clearer product model for continued learning.
V3 - Designing for Return Motivation
How the Design Evolved
The return page evolved alongside VideoTutor's product strategy and feature expansion. As the product grew beyond K–12 test prep, new capabilities such as tutor personas and exam mode were introduced, while the visual language matured into a cleaner, more universal experience.
Switch Tutor
Upload file
Daily Challenge
Fun Facts
03
In-class experience
The old classroom was built around a finished video. Phase 3 needed to support live tutoring, lesson changes, and future learning tools without making the learner manage a complicated interface.
My design direction was to make the lesson the stable center of the experience. The tutor, controls, and system feedback would appear only when they helped the learner ask, decide, or continue.
Product shift
From generated video to live learning session
UX architecture
Lesson, tutor, and tools working as one experience
Business value
A scalable classroom for broader subjects and modes
Previous Iterations
See the Phase 2 exploration
V0 - MVP (before I join)
The classroom was a video page with extra features around it. Learning still happened inside a fixed video player.
V1 - Early Beta Launch
Retention-focused video experience
The beta added recommendations and return paths, but learning still depended on video consumption.
V2 - Product Evolution
Interactive lesson exploration
I explored a lesson format where the content, question, tutor, and learner input could respond to each other in the same learning flow.
The problem was NOT how to add AI chat to a video player.
The classroom needed a new operating model: the lesson had to stay central, tutor support had to appear in context, and every AI-driven change had to feel understandable to the learner.
01
The explanation stays visually dominant, so learners can follow the concept without managing competing panels.
02
Bring tutor support into context
The tutor opens beside the lesson, keeping the question, answer, and visual explanation connected.
03
When the AI listens, updates, or changes the lesson, the interface shows what is happening and why.
Thank you for viewing my work :)
VideoTutor collection · Overview
Phase 3 Case Study →
Back to top ↑
Work in progress...
Still turning months of chaos into a coherent story.