AI voice assistant

Affrima: Self affirming AI tool

Lab

Eheart Lab

ROLE

Research Assistant Technical Lead

Working under

Kimi Wenzel

YEAR

2024

Project description

Affirma is a voice technology designed to promote subjective well-being through self-affirmation, savoring, and gratitude. The project leverages voice-based interactions to help users engage in daily reflections, fostering emotional resilience and positivity.

Timeline

Feb. 2024 - July 2024

Background

Affirma is a project under Eheart Lab at HCII at CMU.

I was personally responsible for both the technical implementation and refining the user experience.

Please view the rest on a tablet or desk top.

Process

Technical Implementation

I developed core features using Amazon Alexa, GPT-4, and AWS, ensuring seamless voice interactions for personalized well-being exercises. This included creating the backend architecture that facilitated dynamic conversations and customized responses based on user input.

User Flow Refinement

Working closely with the team, I helped optimize the flow of voice interactions, ensuring that users could engage with the platform effortlessly. This included refining onboarding processes and designing check-ins that fit naturally into users' daily routines.

Features Implemented

Onboarding & Personalization

I designed a user-friendly onboarding process that gathers personal values and preferences, tailoring the experience for each individual.

Morning & Evening Check-ins

Developed the flow for daily check-ins, allowing users to reflect on their values and focus on positive experiences throughout the day. These reflections were powered by GPT-4 to ensure nuanced, engaging interactions.

Gratitude and Savoring Exercises

Integrated specific techniques like memory building, sharing experiences, and gratitude into the daily reflections, encouraging users to maintain a positive mindset and improve overall well-being.