STEPS

I explored academic papers, workbooks, PST worksheets, and a coaching audio recording to understand the process and users. I also interviewed clinical experts to learn about common challenges people with TBI face during training.

(1) Cognitive patterns: Users tend to rush into solutions and struggle with setting realistic, measurable goals. (2) Emotional support: To reduce overwhelm, training includes the option to “do nothing” and encourages focusing on one task at a time. (3) Training dynamics: The program aims to build independence. Users often need support during planning calls and benefit from weekly progress check-ins.

I explored a fictional chatbot persona to add gamification as stakeholders suggested. Although the idea wasn’t selected, it revealed that users preferred a simpler, clearer tone—easy to understand and use. To ease the transition from paper to digital, I designed features based on the original workbook, including a chat page, milestone tracker, and help section.

We held monthly meetings with stakeholders, clinicians, developers, and experienced users to improve the design. Key takeaways were: users found the interface confusing, some preferred real coaches instead of cartoon characters, and feedback showed some dialogues needed revision. The team suggested creating a conversation script to guide development.

We evolved from a rule-based system to a hybrid model combining structured steps with AI-driven emotional support to help users feel understood. To support PST’s focus on independence, AI was limited to emotional support and conversation organization, avoiding direct answers. AI-generated ideas were replaced with a library of user-contributed solutions to gently aid brainstorming. Conversation summaries were added to improve progress review.

Designing the PST conversation flow was challenging due to limited references and confidentiality. To address this, I researched similar studies and analyzed notes, then broke each step into smaller prompts to ensure clarity and therapeutic value. To reduce fatigue, I kept text concise and added step breaks. For example of problem assessment, I split the step into three parts and used button-based responses to help users ease into the process without feeling overwhelmed.

To reduce cognitive load, I focused on essential features and kept the design minimal to avoid confusion during testing. The interface supports accessibility with a voice assistant and accent options. I used high-contrast colors—light brown and dark green—to convey warmth and stability, and chose Radio Canada and Inter for readability and tone. Warm-toned icons were added to maintain a cohesive and approachable design.

Clinical experts will conduct formal user testing to evaluate the app’s effectiveness against traditional coaching. To reach a wider audience, the app will also support multiple languages.