A controller is a UI paradigm, like the keyboard & mouse "desktop" paradigm. Serving a userbase that can't use a standard UI will lead to invention of new ones. This could easily lead to mass market products/features that people will want regardless of disabilities.
Particularly interesting (to me) are paradigms being developed for the blind, because they can't rely on displays. This is the same sort of problem as amazon echo, driving mode phones UIs and other products that need paradigms.
The key though, the key is that blind people are a "captive" market. They can't just use the display UI every time the voice UI sucks. This gives developer the opportunity to keep improving it until the number of use cases where voice UIs work better than display gets big enough.