We have built a skill which controls audio playback, volume and routing around a home. Customers purchase our hardware and they can route audio from any device that they have all around their home. Commonly devices like Sonos and services like Spotify are commonly output from our hardware to in-wall or ceiling speakers.
The problem is this:
If you are listening to music in your kitchen and the volume is high, you can tell the Echo via our skill to reduce the volume to a more pleasing level.
The problem is that every echo device (except for the Fire TV Alexa remote) will almost 100% fail to hear the command as it will try and listen to the song lyrics instead!
Now if you listen to a music service like spotify directly from an Echo device, then when you say "alexa" or any other official wake word, the audio from the spotify stream is reduced. We know from our own testing this improves listening from the Echo device almost back to normal.
My question is this: We are unable to find a way to execute a command to our hardware when the user wakes the Alexa with their voice. As soon as the word is spoken we can issue a temporary volume down command so the Echo device has the best possible chance of hearing. Without it, people are blaming us or the Echo hardware for not being able to hear correctly!