Paula
A voice-based companion app that supports elderly people with dementia through gentle reminders and empathetic interaction, inspired by Knapp’s Relationship Model.
Project Type: UI UX
Role: Designer
Duration: Q4 2024
A voice-based companion app that supports elderly people with dementia through gentle reminders and empathetic interaction, inspired by Knapp’s Relationship Model.
Project Type: UI UX
Role: Designer
Duration: Q4 2024
Paula is a concept for a voice-based mobile assistant designed to support elderly individuals—especially those living with dementia—through companionship, reminders, and meaningful conversation. The goal was to create a product that goes beyond utility and builds a sense of trust, care, and connection between user and interface. Rooted in Knapp’s Relationship Development Model, Paula explores how emotional bonding can unfold between humans and machines, and how design can nurture these stages through consistent interaction, memory reinforcement, and comfort.
My personal goal was to push past surface-level utility and explore how digital products can become emotionally intelligent companions. I wanted to create something that didn’t just function well, but felt good to use—especially for those who often feel left behind by mainstream tech. I also challenged myself to design for vulnerability: how can a product earn trust when its user might be forgetful, overwhelmed, or lonely?
We began our research by exploring the daily realities of elderly individuals—especially those coping with memory decline. Through interviews with caretakers and desk research on aging, cognitive decline, and assistive technologies, we uncovered patterns of emotional needs and digital gaps. Many older users feel overwhelmed by rigid tech systems or forget why they opened an app in the first place. What they seek isn’t another tool—it’s reassurance, presence, and gentle guidance. Our professor introduced us to Knapp’s Relationship Development Model, typically used in interpersonal communication, as a framework to map out how a relationship could unfold between a human and an AI assistant. This shifted our thinking from “task completion” to “emotional connection,” and helped us prioritize trust-building over complex functionality.
Elderly individuals—especially those experiencing dementia—often lack consistent, empathetic support in their daily routines. While caretakers try to help, many older users are left alone for long periods, with technology that feels cold, overwhelming, or inaccessible. There’s a need for a companion that feels human, understands their routines, and responds with patience and clarity.
This project gave many moral thoughts, some of which were...
We developed two user roles: seniors and caretakers. The senior user’s journey focuses on daily reminders, AI interaction, and emotional support, while caretakers gain access to vitals, appointments, and doctor notes.
Based on our research and Knapp’s relationship stages, we defined the goal of Paula as more than just a functional assistant—it needed to be a companion that slowly forms a bond, much like a person would. From “Initiating” to “Intensifying,” Paula had to feel trustworthy, personal, and never overwhelming.
We focused on early-stage relationship cues: using the user’s name, keeping language soft and familiar, and avoiding robotic command structures. At the same time, we mapped moments where Paula could “check in” gently, not just with reminders, but also with conversations that made the user feel seen.
We narrowed our focus to three pillars:
Once our focus was clear, we mapped out the user flow based on both technical needs and emotional touchpoints. Inspired by Knapp’s stages, we visualized Paula’s interaction path as a relationship arc—from first introductions to daily routines and eventual emotional reliance.
Each stage was supported by micro-interactions: Paula would ask short questions, react empathetically to emotions, and evolve her responses based on previous conversations. We also explored voice tone options, icon styles, and patterns that would feel familiar rather than futuristic.
The user journey was designed to reflect a natural emotional progression, inspired by Knapp’s relationship model. It begins with a gentle onboarding process where Paula introduces herself and learns about the user’s preferences and needs. From there, the experience flows into daily use: Paula reminds the user of their medication, upcoming appointments, or simply offers a kind check-in. The user can also start conversations, ask questions, or listen to calming content like music or bedtime stories. Over time, these interactions become more personalized and familiar, mirroring how relationships evolve. Paula moves from being just a tool to a trusted companion—integrated into the user’s daily routine and emotional space.
We crafted a simple, branching architecture with two user types in mind:
This division ensured that Paula’s front-facing experience stayed gentle and undistracting, while still offering powerful tools in the background for family members.
We sketched a variety of possible interfaces and flows—some resembling traditional chat assistants, others taking on a more visual avatar approach. After multiple iterations, we landed on a voice-first, face-like interface: Paula had a soft, abstract visual presence that blinked, smiled, and responded subtly. This allowed us to avoid the "uncanny valley" while giving users something to connect with. It also left room for future expansion with voice-based gestures, facial expressions, or even a physical device version.
We translated the wireframes into a mid-fidelity prototype to test core flows—setting reminders, receiving voice prompts, and browsing calendar events. User feedback confirmed that the interactions felt intuitive and emotionally supportive. The voice-first approach helped establish trust early on, aligning with Knapp’s bonding stage.
The wireframes prioritized clarity and comfort. Key functions like medication tracking, conversation prompts, and calendar access were grouped for ease of use. We emphasized simplicity in layout and tested various versions with users before finalizing the structure.
The visual style of Paula was built around softness and clarity. Rounded buttons, warm pastels, and simple icons work together to make the app feel approachable without being childish. Typography is large and legible, with contrast tailored for aging eyes. Every detail supports trust and ease, so that the interface feels like a helpful companion rather than a complicated tool.
The final UI ties together emotional tone and functionality. The layout emphasizes simple flows—like confirming a medication reminder or checking today’s appointments—with high-contrast buttons and minimal screens.
Each component builds trust and consistency, helping users navigate intuitively even with memory challenges. From onboarding to daily interaction, Paula’s design remains quietly supportive.
Our final delivery consisted of a working high-fidelity prototype tailored to the specific needs of elderly users with memory challenges. Each screen was designed with clarity and emotional sensitivity, following both accessibility standards and principles from Knapp’s relationship model. The app’s tone remains steady and comforting, prioritizing clarity and support over complexity.
This project taught me how to translate abstract emotional stages into real interaction patterns. Designing for trust, vulnerability, and reassurance pushed me to consider not only usability but also emotional resilience. It strengthened my understanding of how design can support real-world cognitive and emotional limitations.
Tools used:
A user-friendly marketplace app designed to streamline peer-to-peer product exchange within the university community.
View project
A music streaming app with integrated social features that encourage sharing, discovery, and meaningful interaction between users.
View project
An interface extension for NextCloud that introduces polls within chat, empowering users to make collaborative decisions with ease.
View project
