I have a pretty sophisticated setup. I often end up in unique ‘corner cases’ that nobody else notices.
We use multiple Google Home (Nest) displays/speakers for voice commands. There’s pretty much one in every room. Most of my Home Assistant devices are ‘templated’. That means the original device is hidden and only the template is exposed to Google. I also use the alias feature to add additional names for use with voice assistant routines. This is because Google has a mind of its own with words like light, fan, vanity, living, etc. For a voice prompt like ‘living dim’, my Google command is ‘set 1!v!n9 to dim’. Oh yeah, my templates also have named preset values.
All my Google automations are now in the web-based script format. Because there’s no option to respond on the same device, they’re all silent. The home automation is set up under a separate dummy Google account with each family member invited using their own Google account.
Everything worked fine until a few days ago. I just checked my Android phone and it was set to use Gemini. I forced it back to the old Assistant. The commands work on my phone using either text or speech. However, when I use a speaker, I can see the correct phrase on the display, but nothing happens. There’s no indication of error. The processing delay seems quicker than I’d expect. When I view Google Home ‘activity’, only the commands issued through my phone show up.
Toggling [Activate] on a routine does not help. Neither does delete and recreate as new.
Normally, I would just wait for someone else to waste their time reporting to Google. Has anyone else noticed a change?