Curating Befreed’s AI-powered content digestion and discovery experience
Project at a Glance
Befreed, a mobile reading app that breaks down the entire book in to bite-size content with AI. Product Team is seeking for design solutions to improve the user retention rate and attract more users.
Through user research, three key insights were uncovered, which informed the ideation of potential design solutions for Befreed. Based on stakeholder reviews, three design directions were prioritized for further development. These directions were then refined and validated through continuous feedback from both users and stakeholders.
Ultimately, I translated the final solution into high-fidelity mockups, leveraging components from the new design system in Figma to ensure consistency and usability.

Project Overview
Context
Befreed is an early-stage mobile reading App with 40,000 users that helps time-constrained readers learn efficiently by transforming books into bite-sized knowledge using AI technology. As the UX designer on the team, I led the end-to-end design to improve overall user experience and seamlessly integrated AI into the reading process, contributing to a more intuitive and engaging AI-optimized reading experience.
Outcome
A strategic product roadmap was established by conducting comprehensive user research to identify target audiences and uncover product improvement opportunities.
Average user retention time was increased from 3 to 6 minutes and 90% of users were satisfied with new design during the user testing session.
Product visibility was boosted by showcasing design solutions to 200+ industry participants in cocreate conference as the lead incubator project owner.
Design Process
Product auditing
Current State
The product coverts books into 5-20 minute AI-optimized content with 4 reading styles. The user flow starts on the Discover Page where users can browse and select a book. They can then view detailed information on Book Detail Pages to decide whether to read, and finally choose a reading style to begin.

Problems to be sloved
Multiple usability issues lead to user confusion and disrupt the overall reading flow.
Content lack of variety that results in poor user engagement.
Latent AI capabilities limit the potential to enhance the learning experience.

Research
User Interviews
We decided to conduct qualitative research to collect in-depth user insights and identify key pain points, rather than rely on quantitative methods, which could lead to misleading results due to the small sample size.
I wrote a Research Plan to design a semi-structured user interview, after conducted 16 user interview sessions, we collected more than 60 user feedbacks and used Affinity Map to cluster them in to 10 user insights.

Key User Insights
We prioritized 3 key user insights that resonated with users the most to explore the potential solutions.



Design Directions
In the stakeholder review, we shared key user insights with high-level solutions and actionable next steps, and ranked each solution based on design cost and user impact.

Ideation
Brainstorm
By utilizing storyboards, Crazy 8s, and hosting multiple ideation workshops, we began to brainstorm as many ideas as possible.

In our research, we uncovered two user insights that helped shape the design:
Readers expect more detailed content to deepen their learning.
To support deeper learning without disrupting the reading flow, we introduced AI-powered hyperlinks that offer contextual information tailored to each subject. This helps clarify complex ideas, reduce confusion, make learning more accessible.

Readers want to apply what they learn to real-life situations.
Initially, we explored features like summary screens and pop-up quizzes to improve knowledge retention. However, user testing revealed these were often intrusive and distractive. As a result, we grounded our design in cognitive learning principles. The revised solution uses AI to surface key concepts from books with real-life examples, and reflective prompts—guiding users through deeper understanding while reinforcing application through subtle, context-aware support.

Iteration
We put ourselves in the users’ shoes, prioritizing intuitive interactions and refining UI components to arrive at high-fidelity mockups. Three final options to check the contextual information and key concept were selected during our design critique session.
Option1: Swipe to second page to check
Option2: Click the highlight texts in the script page
Option3: A toggle to open the highlight texts in the script page



Next, we curated a user testing session with 3 target users to test the following tasks:
Task A: Find a highlighted term, access its contextual information, and ask a question using the AI chatbot.
Task B: Find the key concepts, try to apply one of the suggested learning outcomes.
Task C: Rate the overall experience and share what part helped you learn the best.
After testing and evaluating all three options, Option 3 demonstrated the highest usability and learning impact, leading us to advance it as our final design direction.

Final Deliverables
Prootyping
Book Detail Page
Offers a clear content breakdown to support a more informed and focused pre-reading experience.
Enables users to explore AI-curated reading recommendations, increasing the user engagement and enhancing content discovery.


Swipe AI
Provides contextual information linked directly to highlighted text, helping users better understand complex ideas while reading.

Surfaces key concepts and real-life applications of the reading content, making the knowledge more applicable and relevant to users’ learning goals.

Mind Map
Encourages users to interact with diverse book galaxies, promoting exploration of interconnected content in an engaging, visual way.

Learnings
Looking back, one of the key lessons I learned during the project was how to make stakeholder reviews more effective. I started presenting ideas with clear design notes to help stakeholders better understand with the rationale behind each concept. I also restructured the meetings to include a dedicated design Q&A session to ensure a clear alignment on next steps. Through this process, I quickly adapted to the fast-paced startup design environment and learned how to drive impact within tight timelines.
“I had the pleasure of mentoring and working with Rachel during our design cohort, and she truly stood out as a talented and reliable designer. She has a great eye for visual design and consistently brought strong creative thinking to the table.
What impressed me most was her leadership — she took initiative, kept her team on track, and always approached challenges with curiosity and a positive attitude. From start to finish, Rachel was all-in, and it showed in the quality of her work and her impact on the team.”
Jacky Cheung
UX Designer | Amazon
Curating Befreed’s AI-powered content digestion and discovery experience
Project at a Glance
Befreed, a mobile reading app that breaks down the entire book in to bite-size content with AI. Product Team is seeking for design solutions to improve the user retention rate and attract more users.
Through user research, three key insights were uncovered, which informed the ideation of potential design solutions for Befreed. Based on stakeholder reviews, three design directions were prioritized for further development. These directions were then refined and validated through continuous feedback from both users and stakeholders.
Ultimately, I translated the final solution into high-fidelity mockups, leveraging components from the new design system in Figma to ensure consistency and usability.


Project Overview
Context
Befreed is an early-stage mobile reading App with 40,000 users that helps time-constrained readers learn efficiently by transforming books into bite-sized knowledge using AI technology. As the UX designer on the team, I led the end-to-end design to improve overall user experience and seamlessly integrated AI into the reading process, contributing to a more intuitive and engaging AI-optimized reading experience.
Outcome
A strategic product roadmap was established by conducting comprehensive user research to identify target audiences and uncover product improvement opportunities.
Average user retention time was increased from 3 to 6 minutes and 90% of users were satisfied with new design during the user testing session.
Product visibility was boosted by showcasing design solutions to 200+ industry participants in cocreate conference as the lead incubator project owner.
Design Process
Product auditing
Current State
The product coverts books into 5-20 minute AI-optimized content with 4 reading styles. The user flow starts on the Discover Page where users can browse and select a book. They can then view detailed information on Book Detail Pages to decide whether to read, and finally choose a reading style to begin.


Problems to be sloved
Multiple usability issues lead to user confusion and disrupt the overall reading flow.
Content lack of variety that results in poor user engagement.
Latent AI capabilities limit the potential to enhance the learning experience.


Research
User Interviews
We decided to conduct qualitative research to collect in-depth user insights and identify key pain points, rather than rely on quantitative methods, which could lead to misleading results due to the small sample size.
I wrote a Research Plan to design a semi-structured user interview, after conducted 16 user interview sessions, we collected more than 60 user feedbacks and used Affinity Map to cluster them in to 10 user insights.


Key User Insights
We prioritized 3 key user insights that resonated with users the most to explore the potential solutions.






Design Directions
In the stakeholder review, we shared key user insights with high-level solutions and actionable next steps, and ranked each solution based on design cost and user impact.


Ideation
Brainstorm
By utilizing storyboards, Crazy 8s, and hosting multiple ideation workshops, we began to brainstorm as many ideas as possible.


In our research, we uncovered two user insights that helped shape the design:
Readers expect more detailed content to deepen their learning.
To support deeper learning without disrupting the reading flow, we introduced AI-powered hyperlinks that offer contextual information tailored to each subject. This helps clarify complex ideas, reduce confusion, make learning more accessible.


Readers want to apply what they learn to real-life situations.
Initially, we explored features like summary screens and pop-up quizzes to improve knowledge retention. However, user testing revealed these were often intrusive and distractive. As a result, we grounded our design in cognitive learning principles. The revised solution uses AI to surface key concepts from books with real-life examples, and reflective prompts—guiding users through deeper understanding while reinforcing application through subtle, context-aware support.


Iteration
We put ourselves in the users’ shoes, prioritizing intuitive interactions and refining UI components to arrive at high-fidelity mockups. Three final options to check the contextual information and key concept were selected during our design critique session.
Option1: Swipe to second page to check
Option2: Click the highlight texts in the script page
Option3: A toggle to open the highlight texts in the script page






Next, we curated a user testing session with 3 target users to test the following tasks:
Task A: Find a highlighted term, access its contextual information, and ask a question using the AI chatbot.
Task B: Find the key concepts, try to apply one of the suggested learning outcomes.
Task C: Rate the overall experience and share what part helped you learn the best.
After testing and evaluating all three options, Option 3 demonstrated the highest usability and learning impact, leading us to advance it as our final design direction.


Final Deliverables
Prootyping
Book Detail Page
Offers a clear content breakdown to support a more informed and focused pre-reading experience.
Enables users to explore AI-curated reading recommendations, increasing the user engagement and enhancing content discovery.




Swipe AI
Provides contextual information linked directly to highlighted text, helping users better understand complex ideas while reading.


Surfaces key concepts and real-life applications of the reading content, making the knowledge more applicable and relevant to users’ learning goals.


Mind Map
Encourages users to interact with diverse book galaxies, promoting exploration of interconnected content in an engaging, visual way.


Learnings
Looking back, one of the key lessons I learned during the project was how to make stakeholder reviews more effective. I started presenting ideas with clear design notes to help stakeholders better understand with the rationale behind each concept. I also restructured the meetings to include a dedicated design Q&A session to ensure a clear alignment on next steps. Through this process, I quickly adapted to the fast-paced startup design environment and learned how to drive impact within tight timelines.
“I had the pleasure of mentoring and working with Rachel during our design cohort, and she truly stood out as a talented and reliable designer. She has a great eye for visual design and consistently brought strong creative thinking to the table.
What impressed me most was her leadership — she took initiative, kept her team on track, and always approached challenges with curiosity and a positive attitude. From start to finish, Rachel was all-in, and it showed in the quality of her work and her impact on the team.”
Jacky Cheung
UX Designer | Amazon