This project, in collaboration with TIRR Memorial Hermann, focused on developing an e-health product to help individuals with traumatic brain injury improve their problem-solving skills and cognitive abilities.I designed a chatbot app that integrates the original phone-based and paper-based training process, making it clearer and easier to follow.
I was responsible for transforming the traditional training process into a digital format, creating wireframes and prototypes, and designing AI-driven conversations for a more natural user experience during problem-solving training.
The traditional process is time-consuming and depends on professional staff for coaching. It also requires scheduling, which limits flexibility. When people use a workbook for training on their own, they often lack sufficient guidance. Additionally, there is a need for a system to better organize and track training sessions.
The stakeholders suggested creating a gamification-like app to make the process more engaging and less monotonous for users. I proposed two ideas: one was a chatbot app with a game-like conversational tone, and the other was a survey-like app featuring different game-like interfaces at each step of the training process.
The stakeholder preferred the chatbot version because it offers more interactions with users, helps manage their impulsivity, and prevents them from jumping straight to solutions.
I researched papers on problem-solving training and reviewed materials provided by the stakeholder, including a workbook, sample worksheets from individuals in training, and an audio recording of a coach explaining the process. To gain deeper insights, I also interviewed a coach to better understand the common challenges individuals encounter during training.
I divided the design into two parts: the user interface and the conversation design. At first, I focused on making it easy for users to record and review their training progress. For the gamification aspect, I suggested using a fictional, cartoon-like character to make the conversation more engaging during training.
My inspiration for creating the persona came from the movie *Finding Nemo*. One of the memorable characters, Dory, has a short-term memory deficit, which parallels some of the challenges individuals with cognitive impairments face. This led me to develop Mr. Ray as a persona, drawing on Dory's traits to make the character both interesting and relatable.
The design features a chat page, menu, milestone page, history page, PST info page, and help page. With my background in healthcare, I understand that people often feel apprehensive about transitioning from paper to digital formats and prefer familiar structures or additional guidance. To address this, I incorporated more guidance based on their original worksheets to make the shift smoother.
After discussions with stakeholders, I realized that a low-fidelity design would be more effective for feedback. I also decided individuals should focus on the conversation rather than the process status, so I moved the status feature below the message area.
%20icon.jpg)
I had a meeting with stakeholders, clinical professionals, developers, other design team, and individuals with prior training experience to discuss their preferences and concerns regarding the design.
I iterated the chatbot design from a rule-based system to a hybrid approach, combining both rule-based and AI-powered elements. The rule-based system maintains the structured steps familiar to users, while the AI component adds emotional responses to create a more natural, empathetic interaction. This approach not only helps users feel understood but also offers personalized suggestions for goal setting and plan development.
I broke down each step into smaller, more detailed chunks to ensure effectiveness and clear goals. For example, in assessing a problem, I divided it into three parts: (1) defining the problem, (2) expressing feelings and context, and (3) evaluating its importance. The first part helps stakeholders identify common issues without accessing confidential information. The second allows individuals to share feelings without judgment, while the third reinforces the purpose and importance of the training.
I focused on simplifying features and ensuring buttons and icons were visible and easy to interact with, leading to a one-page for the web version. I also replaced the coach icon with images of real people, as users prefer interacting with real individuals.
During onboarding, users can select their coach, and a brief audio greeting is provided. This allows users to choose a voice they feel comfortable with.
On the homepage, users can review their weekly tasks and see reminders or advice from their coach. The large user and coach icons allow for easy access to settings.
To provide users with direct guidance, a prompt in the message area always informs them of the next steps to continue the process. Additionally, I designed many questions with button-based answer selections to help reduce mental workload and make the process easier to navigate.
The archive allows users to review their previous training and track the percentage of the process they’ve completed.
I divided the screen into sections to fit all features on one page, making it easier for users to view all elements at once and navigate without fear of getting lost.
Larger buttons enhance accessibility for users with visual or physical impairments.
