Everything voice, feature by feature.

One device and one workspace, three behaviours: capture what happens around you, speak what you need to write, and ask what you need to remember.