The problem
A companion is only useful if people can understand what it is communicating. Expressions and reactions need to be legible, relevant, and not disruptive.
AI & SOFTWARE
Sori Companion.
A companion concept focused on understandable facial expressions and activity-driven reactions, rather than a full 3D character or movable robot.
PrototypeRole: Product and interaction designer
PROJECT / SORI-COMPANION
How I'm approaching it
- Use a persistent on-screen avatar as the initial form.
- Connect reactions to explicit behavior and user activity signals.
- Keep the first version achievable without requiring a 3D character engine.
- Treat attention and activity sensing conservatively, with clear controls and privacy boundaries.
What this work is teaching me
Behavior can feel playful without being intrusive. An understandable set of states is a better starting point than a large list of random animations.
What's next
Define the behavior-state table and accessibility/privacy rules before expanding the companion's reactions.
Descriptions reflect current portfolio notes. Update them as the project evolves.
HAVE SOMETHING IN MIND?
Good work gets better when people connect.
Open to thoughtful feedback, learning alongside other builders, and meaningful collaborations.