Physical AI for service design

This was a fast-paced, experimental concept design project in partnership with Strange VC and two clients:

  1. Archetype AI, focused on physical AI
  2. Roche, in healthcare

Archetype AI came with the request to map its multimodal AI capabilities to Roche's healthcare use cases.

The team was small—an engineer, a PM, and me. My role was to conduct the research, including creating the service blueprint and analyzing when to use AI and when not to use AI. The three of us collectively shared ideas and created the final artifacts.

  • Role: UX Researcher, conducted research and created service blueprint
  • Team: 1 engineer, 1 PM, 1 designer

Problem

The problems we wanted to solve were:

  1. Errors in the hospital-to-lab experience, with the goal of minimizing errors and reducing waste.
  2. How to be technology-driven but not tech-centered: finding the right intervention point to maximize the model's capabilities without forcing the technology into the experience.

Outcome

Physical AI healthcare interaction design
Hardware artifacts: base station, tracker base, and tracker module
Proposed user flow with 8 interaction steps

Storyboard

When, where, and why AI should exist

Storyboards helped stakeholders imagine the role of the system in context. The focus was less on a screen and more on when, where, and why AI should exist in the status quo.

My Role & Research Process