Back to Work
Applied Research + AI Interaction Design

Alexis: Qualitative AI Agent

Conversational DesignPrompt EngineeringBehavioral ScienceUser Testing

What It Is

Alexis is a qualitative interview agent designed to study emotional patterns in everyday digital behavior. It's a conversational system that asks context-aware questions, follows up naturally, and captures rich qualitative data without overwhelming the user.

Why I Built It

Traditional qualitative interviews are slow and expensive. I wanted to explore if an AI agent could collect rich emotional data while keeping users comfortable, leveraging the "low friction" nature of automated interactions to help people open up.

Designing the Interaction

Conversation Mapping

Mapped intricate conversation flows for emotional, behavioral, and reflective questioning paths.

Tone & Pacing

Designed the language style and pacing to feel natural, refining prompts to avoid "scripted" robotic responses.

Contextual Backend

Built a backend that stores conversational context, allowing the model to guide the interview intelligently.

Vulnerability Analysis

Identified patterns in how users share vulnerable information with AI, testing where interactions felt natural vs. mechanical.

Why It Matters

This project sits at the intersection of UX research, behavioral science, and AI. Alexis isn't just a bot; it's a validated research tool. It demonstrates my ability to balance trust, emotional safety, and research validity while building functional technical systems.