A few years ago, I saw a demo of music being generated based on the emotion and tone of the content being spoken as TTS. It was incredible to see how a bland TTS service could come alive just by overlaying music.
Those weird robotic phrases seems to blurred by the background music and the words themselves became more convincing. It’s amazing how much layering emotion inducing music can add to the content of words.
There are already engines out there that can read the tone of language and provide that information back. There are also playlist suggesters for complimenting moods.
The low hanging fruit would be to combine these two — Mood as a Service — that could produce music or imagery to compliment the tone of a spoken or written piece.