Kuri, the AI personal assistant, is a web app that caters to a wide variety of users. Our product creates a way for users to streamline their workflow with the AI's assistance by integrating with other applications, and providing information in a clear and concise way that is catered to the user's individual wants and needs.
UX Designer
Figma, FigJam, Google Docs, Teams, and Discord
February - April (12 Weeks)
Team Lead
Interaction Designer
UX Designer
Interaction Designer
UX Designer
Goal-Directed Design (GDD) is an approach on design thinking and how to design products that focuses on how to create for specific user types and their goals. While features and functionality is important, GDD focuses on meeting the needs of the user above all else.
This design process is broken down into several phases that help designers focus on research and analysis (Research) , defining the goals of users (Modeling/Requirements), developing concepts of the product (Frameworks), and refining the final product through usability testing (Refinements).
Using this process helps designers focus on the goals that their users might have, using research and design iterations to develop and improve their product.
At our team's Kickoff meeting, we collectively decided that we wanted to focus on new technology. At the time, AI products were new, with ChatGPT released just before starting our project. We thought it would be exciting to work with the idea of Artificial Intelligence, and learn with the rise of AI.
We began by making general assumptions about what we wanted to accomplish, even with little knowledge of AI at the time. This would give us a general idea of the direction that we wanted to head. Our Kickoff Meeting concluded and we were left with our initial Problem Statement
The current state of the AI tools has focused primarily on providing conversational answers to users. What existing products/services fail to address is the ability provide responses and guidance that meet specific preferences of their users . Our product/service will address this gap by utilizing artificial intelligence to create a personalized guide that can help with areas specific to your life
To begin, our team conducted a lot of in-depth research. At the time of starting this project, AI was a relatively new field and we wanted to better understand how it worked, how people were using it, and what their feelings were towards it. We also wanted to study people's search behaviors and how they interacted with search engines, and their methods for saving and storing information.
We believe it to be incredibly important to research into the good and bad sides of what would be our products competitors. We took several hours and looked into other AI apps such as ChatGPT, DALL-E 2, Midjourney, and later Bing and Bard once they had released. We spent time within these apps exploring both the AI interaction, and the features that were offered by these different programs, such as their storage and subscription services.
We also began looking into the differences between various search engines, looking specifically at Google, Yahoo, Bing, and DuckDuckGo. By understanding the differences with these, we could craft some of these features into our program to make sure that users would feel comfortable and familiar with what we would be making.
Our literature review focused on a variety of sources that would provide us factual information to back up the assumptions we had created in our Kickoff Meeting. While a majority of our research focused on bettering our understanding of AI, we spent a lot of time looking into people's perceptions of AI, regardless of the facts about how it works. This would help us better understand people's hesitance or excitement towards AI, and how we could create a product that would be inviting to all types of users.
After completing our initial research, we reached out to friends and family members that we believed would be potential users of the product that we were creating and conducted ethnographic interviews that would give us valuable insight into a user's opinion of our problem space. For all of our interviews, we set up Teams meetings where we asked them general questions in three major areas: their search engine behaviors, their perception of AI, and their methods of saving and storing information found on the internet.
For one of our final interviews, we had the opportunity to talk with Ryan Webber, a Computer Science Graduate Student at Kennesaw State University, who served as our Subject Matter Expert. He was able to provide us a deeper understanding of the backend of Artificial Intelligence, how it works, and how he could imagine it developing in the future.
He explained that right now, we are in an area of infatuation with AI, and that people were adding it to their products just to jump on the hype of it. He expressed some concern and warning that we needed to make sure our product did not fall within the same category, so our team made sure to put more research into how AI could be truly beneficial for our users.
At the end of each interview, our team completed an Affinity Map. This was a culmination of all the notes that we took individually, and collected at the end to make sure we were all on the same page with what we had learned. By doing this, we also began to note key behavior patterns that would be useful in the next phase.
Once everything had been grouped by behavior, we completed a summary that would serve as a quick recap of all the information that we identified.
With our research gathered and organized, and our interviews completed; our team moved into the Modeling Phase. Here we began to conceptualize our research findings, analyzing it into behaviors and patterns that we could then formulate into Personas. These personas would be the ideal users that we identified from our research, and would become the foundation of our design process.
Once our interviews were completed, we began listing many of the areas we focused on with our questions. We formatted these into common behaviors that we noticed amongst all our interviews.
We then analyzed each individual affinity map, and charted their responses on behavior continuums. Once everyone was plotted, our team began looking for patterns and trends amongst them. From these patterns we were able to identify two main user types that would be the foundation for our personas.
To summarize all of our research, our team identified two main user types that formed from our interviews and behavior variables. After analyzing them further, and discussing our results as a team, we solidified that we would have one primary persona who we could design the app for with all it's features. This user would be more knowledgeable about the app and be like a returning user. Our secondary persona would cater to the user type that we identified that was more skeptical about AI, but was willing to use it as a more creative outlet.
So after dividing and creating small biographies for our personas, I led the rest of my team in creating and detailing the goals of each persona, focusing on what they wanted to accomplish with our app (End Goals), and what they wanted to accomplish on a larger scale (Life Goals). That way we could make sure that our app would cater to their needs specifically.
Daniel Price is a new product manager for a startup in NYC, currently adjusting to his leadership role and struggling to organize his team. He prioritizes efficiency and relies heavily on technology to stay organized. Despite his busy lifestyle, he seeks new technologies to assist him in creating a schedule that caters to his needs.
A graduate student from Miami, Kim prioritizes accuracy and efficiency in her research. She values concise and accurate information. She is conscious of her personal privacy and safety and takes precautions when browsing the internet. Kim's drive to succeed makes her a resourceful learner who is exploring the benefits of technology for her personal hobbies and creative outlets.
With the storytelling aspect of our personas completed, our group then had to theorize what the personas would need and want from the app. This is what the Goal-Directed Design process calls the Requirements Phase. Here, we take our user personas and imagine scenarios in which they would be using the app. From these scenarios we then create lists of features that we would need to include that would satisfy the persona's goals.
With our personas completed, our next step was to craete Context Scenarios that would give life to the activities and operations of our personas. Context scenarios allow us as designers to imagine a day-in-the-life of our persona, what they would be doing, the actions they would need from our product, and other constraints they might come across.
I wrote a context scenario for each of personas that highlighted a day of their lives when using Kuri. By doing this, I was able to highlight specific needs and wants that they might have and putting myself in the shoes of the user.
From the context scenarios, our team was able to pull out specific actions and features that would need to be included. This helped to conceptualize the things that we could design in order to satisfy the needs of our personas.
We split our lists into several sections, highlighting the difference between general requirements, Daniel's requirements, Kim's requirements, and any overlapping requirements.
The Frameworks Phase is where we begin to build out our physical project. We begin this phase by creating features that would satisfy the needs of our personas, and create Scenarios that detail the most used paths through our application. Then, using these as our foundation, we begin to build the wireframes that shape the design of the final product.
To begin, our team performed a Task Analysis. This allowed us to identify all of the features and actions that users would need to work through the app and satisfy their end goals. With our list created, we then organized our Key Path scenario that would be the most used of our app, and focused on using the chatbox. From then we were able to add validation scenarios for other various features we needed to implement. This included the initial onboarding process, as well as the search feature, andd various options that the user would be able to interact with on the chat itself.
Using our key path scenario as a basis, we then began wireframing on Figma. We focused only on the main elements at first: the main chat page, the search panel, the navigation panel, and the settings page. These wireframes were kept very loose and focused more on high-level structure to convey the idea of the product, and the layout of certain elements like the chat bubbles. We would then take these into our first Usability Test, where would would gather feedback from one of our user research participants.
Once our Wireframes were completed, our team moved into the final phase of GDD that we would focus on, the Refinements Phase. This is where our lo-fidelity wireframes became hi-fidelity prototypes that would operate under a fleshed-out design system, and finally end with usability testing.
Before moving into hi-fidelity prototyping, our team worked to create a design system for our project. This would ensure that all of our spacing, colors, typography, and components would remain consistent across Kuri's screens. We split up the sections that we wanted to focus on, and created the style guide for Kuri, basing it off of Apple's Human Interface Guidelines to ensure that we would be following accessibility and web standards.
If given more time, we would have liked to implement more features, that we would explore and better align to Apple's Guidelines. However, for the time constraints of our project, we focused solely on what was necessary for the product, which included the brand colors, spacing, iconography, and typography.
Once our design system was in place, we began prototyping and finalizing the designs that had previously been laid out in our wireframes. The whole process was relatively quick, as the extra work had been placed into our component library and content writing. We spent several weeks designing the intricate component pieces and stories that would be told in our chats, so when it came for designing the screen, everything worked like a drag-and-drop system.
Once we had completed our screens, we were able to perform a usability test with one of the participants from our user research interviews. Due to time constraints, this interview was very impromptu and informal, but we were still able to gather valuable feedback. The interviewee gave us feedback on his overall opinion of the UI design, as well as helped us identify some issues with labeling, that we had not noticed in our time working on it.
Overall, we gathered useful critiques and feedback on our project. However we do recognize that a sample size of one individual is not ideal for proper usability testing, and if given the time, we would have performed several more interviews to solidify our design decisions.
After completing the prototype and the usability tests, our team organized all of our pages, and laid out the different scenarios that we wanted to highlight when presenting our product. Working through the Goal-Directed Design process helped all of us to focus on letting our research guide the product, instead of designing a product that just looks nice.
Kuri, the AI personal assistant, is a web app that caters to a wide variety of users. Our product creates a way for users to streamline their workflow with the AI's assistance by integrating with other applications, and providing information in a clear and concise way that is catered to the user's individual wants and needs.
Final Prototype
Kuri was designed during our Senior capstone class at KSU, giving me and my team the opportunity to revisit the Goal-Directed Design process. This project came almost completely out of our initial research into AI, as we started with an area of study and not a problem to solve. In doing our research, the problem presented itself and allowed us to figure out a way to solve it.
If given the time and opportunity, we would have loved to flesh out more of the customization options, as they were something frequently mentioned in our user research. We wanted to explore more with implementing different media types as well, including audio and visual elements that could be input or output by the system or user.
And finally, while we are happy with the final design of our product, we would have wanted the opportunity to do more usability testing and user research, as it's what our project was created from.
