AI-powered personal assistant designed to answer questions, generate content, collaborate on projects, and automate tasks.
The research phase of our project consisted of a deep dive into the generative AI domain, mainly through literature review, competitive audit, subject matter expert (SME) interview, and user research.
Besides contributing to user research interviews with notes, observations, and eventual moderation, I was tasked with collaborating with two other members of the team in performing a competitive audit of digital products leveraging generative AI technology, such as ChatGPT, Midjourney, and Dall-e.
During our user research interviews, we applied ethnographic methods and participant observation to gather qualitative data and rich description of context. The six research participants from varying ages and backgrounds were mainly asked about their general perception of AI technologies, safety and privacy concerns, and searching/storing/retrieving information or digital content.
After each interview, we synthesized our notes using affinity mapping to track initial patterns and imperatives. Ultimately, our research gave us a better understanding of our potential users’ goals and needs as it relates to the domain of our product.
The modeling phase involved the integration of behavior patterns, motivations, attitudes, opinions, and mental models gathered in user research, and the creation of personas, which are models of represented personalities of our “ideal” or “typical” user.
Our team collaborated on plotting research participants’ data into a visual continuum matrix, identifying clusters and trends out of behavioral variables, and creating our primary and secondary personas, Daniel and Kim.
In Goal-Directed Design, drawing a requirements list is done through context scenarios, which are hypothetical narratives that describe a “day-in-the-life” of our persona so as to speculate details about the many contexts in which they might interact with our product. In addition, our team ideated on solutions while maneuvering technical constraints and addressing business goals.
For our primary persona, the context pointed at a need for organization, categorization, and efficiency in managing content, while for our secondary persona, they focused on customization and privacy. In both instances, users valued an uncomplicated system and familiar product environment.
With user goals and needs in mind, and while considering system constraints, our team started to sketch the user interface through wireframes. This step served to block out the layout, information architecture, and visual design.
Kuri’s interface leverages design principles and best practices such as Jakob’s Law (users prefer familiar systems and interactions) with the use of a search bar to locate specific chats, and Law of Proximity (objects near to each other tend to be grouped together) with the configuration of chat bubbles, content, and settings.
The refinement phase marks the transition between low-fidelity wireframes to a high-fidelity, interactive prototype. For this step, we relied heavily on Figma’s Components and Auto Layout.
In building Kuri’s prototype, I collaborated with another teammate to create a design system following Atomic Design principles, which borrows from chemistry to create categories and hierarchy that help structure interface design components -- atoms, molecules, organisms, templates, and pages.
Despite the challenges of designing experiences for users of novel technology such as generative AI, our team was happy with Kuri as a concept for a product that could compete with other chat-based solutions, especially considering our emphasis on leveraging system components and functionalities that are familiar to most users of information technology.
If given more time or the opportunity to revisit project Kuri, I would perform usability testing of the final prototype with real users. As invaluable as this could have been in validating our assumptions in regards to the overall user experience, we as a team had to deal with the constraints and demands of our capstone class schedule.
