JG
ExpressAI
Empowering Communication Through AI
The Communication Barrier
For many non-verbal autistic individuals and people with physical conditions that affect speech, such as tongue cancer, everyday communication can be a challenge. ExpressAI was created to address this barrier through an AI-powered text-to-speech application designed to help users express themselves clearly and independently. This project was completed at DesiHeart, where I worked as a UX Designer alongside six other designers and a team lead. Together, we followed a structured design sprint process rooted in accessibility, empathy, and user-centered design.
Rethinking Assistive Communication
While assistive communication tools already exist, many fall short in areas like personalization, emotional expression, and ease of use. ExpressAI was designed to rethink this experience by leveraging AI to generate natural-sounding, sentiment-aware speech, predictive text, and adaptive communication features. The goal was not just to help users communicate, but to help them feel heard, understood, and empowered in their daily interactions.
Designing Through a Human Lens
During the early stages of the project, my focus was on developing user personas that grounded the team’s work in real human experiences. I created personas for Alex, a child with autism who relies on visual, routine-driven communication, and Mark, a teen with cerebral palsy who values independence and uses assistive technology to express himself. These personas went beyond surface-level demographics, capturing emotional needs, communication frustrations, reliance on caretakers or adaptive tools, and the desire for autonomy and dignity. A recurring theme emerged: both users didn’t just want to “say” things, they wanted to feel heard in a way that reflected their emotions, personality, and urgency. This work helped shape the foundation for the rest of the team, who were simultaneously conducting competitor analysis and mapping user flows, allowing us to build a shared understanding of both the functional and emotional gaps in existing solutions.
Mapping Moments That Matter
After completing the user personas, I created user journeys to map how both personas would interact with the app in real-life scenarios, focusing on both functional steps and emotional experiences. For Alex, a child with autism, the journey emphasized urgency and predictability, highlighting the need for a simple, low-sensory interface with quick access to emergency communication. Pain points such as visual clutter or delayed alerts showed how easily stress and confusion could escalate in critical moments.
For Mark, a teen with cerebral palsy, the journey focused on building and expressing more complex thoughts using assistive inputs like eye tracking. His experience emphasized independence, efficiency, and dignity, revealing how slow interactions, small touch targets, or unnatural voice outputs could disrupt communication. Together, these journeys reinforced that communication is not just functional, it is deeply emotional and design decisions directly impact a user’s confidence, control, and ability to connect with others.
Designing With Emotion in Mind
As a team, we each created our own mood boards to explore the emotional direction of the app. Rather than focusing only on visuals, we focused on how the app should feel approachable, calm, expressive, and human for all users. After reviewing all of the mood boards together with our project lead and the company owner, we aligned on a design direction that balanced accessibility with emotional depth. This decision influenced everything from layout and color choices to interaction patterns and voice output, ensuring the experience felt supportive rather than clinical.
Building and Testing in a Real-World Context
When building the prototype, the team was divided into smaller groups, each responsible for a different section of the application, including the phrase builder, add custom word, login and account creation, emergency flow, and settings. Myself along with two other designers focused on the “add custom word” experience, ensuring users could easily personalize their communication in a way that felt intuitive and efficient.
We then tested our mockups with a caretaker of a non-verbal autistic individual to understand how the app would perform in a real-life setting. The feedback was valuable. The sentiment-based voice output was seen as a strong improvement over traditional tools, helping communication feel more natural. However, the testing also revealed areas for improvement, especially around navigation speed, clarity, and ease of use during stressful situations.
Iterating Toward a Better Experience
Using this feedback, we refined the designs to better support both users and caretakers. We simplified navigation, reduced cognitive load, and made key features, including emergency communication, faster and easier to access.
We also ensured that emotional expression remained clear without making the interface overwhelming, maintaining a balance between function and feeling.
Key Learnings
Working on ExpressAI deepened my understanding of accessibility and inclusive design. It reinforced the importance of designing with empathy and showed how even small decisions can significantly impact usability. Collaborating with a larger team also strengthened my ability to communicate ideas, iterate quickly, and align on shared goals. Most importantly, this project demonstrated how thoughtful design and AI can meaningfully improve quality of life.
