AI-Powered Recommendations
Milkinside partnered with Crossing Minds to enhance the design of Hai, a unique mobile app offering cross-domain recommendations—from books to food—powered by advanced AI. Central to Hai's functionality was the AI's ability to learn user preferences, making the user-AI interaction a key element of the product's narrative.
Role
Product Experience Designer
With
Milkinside
Client
Hai
AI Training through User Interaction
Challenge
The main challenge was designing an interface that effectively trained the AI based on user inputs, beginning with onboarding. Ensuring users understood and engaged with this training process was crucial for the AI to provide relevant and accurate recommendations across various verticals.
From User Research to Prototype Testing
Approach
To address this, our approach was multi-faceted. We began with comprehensive user research to understand various interaction paradigms, ensuring that we could create an intuitive and effective onboarding experience. This research informed our design decisions, helping us tailor the AI interactions to user behaviors and preferences. Building on this foundation, we developed prototypes that were instrumental in visualizing the user-AI interaction.
These prototypes were not just design tools; they served as crucial elements for user testing, allowing us to gather feedback and refine the experience continuously. The testing phase was iterative, focusing on both the functionality and the narrative aspect of the AI interaction, ensuring that the training process was as engaging as it was educational. This thorough process of research, prototyping, and testing ensured that the final product was not only technically sound but also deeply aligned with user needs and expectations.
A Three Dimensional Discovery Experience
Solution
In redesigning Hai, our objective was to seamlessly integrate discovery into every aspect of the app's user experience. This began with a refined onboarding process, thoughtfully introducing users to the AI's capabilities and ensuring a natural progression into the app's broader functionality. The UI redesign extended beyond basic usability enhancements, incorporating a two-dimensional discovery matrix on the X and Y axes to facilitate an immersive exploration of AI-curated content.
Additionally, a significant focus was placed on the Z space, deliberately crafted for deeper, more intentional AI interactions. This area was designed for users seeking an active discovery experience, offering a contrast to the passive engagement typically found on the app's homepage. This intentional layer in the Z dimension was crucial in creating a holistic discovery journey, allowing users to navigate through and interact with the AI system in a more meaningful way. Balancing the technical intricacies of the AI recommendation engine with clear, accessible user value, the three-dimensional framework I developed not only enhanced the app's intuitive nature but also ensured its adaptability and scalability for future development needs.
Designing Conversational AI Interfaces
Reflection
Working on Hai was a remarkable experience that blended UX design with cutting-edge AI technology. It pushed me to think innovatively about user interactions and their impact on AI training. This project reinforced my belief in the power of user education and conversational interfaces to create not just a product but an experience that resonates and evolves with the user.