Be it Google Assistant SDK or Amazon AVS, these SDKs are mainly used for voice services. If you add an additional layer on top to customize the wake word/initial trigger, won't these systems/setups be something like : Custom wake word processing Engine -> IFTTT Maker or Amazon skills -> Underlying subsystem that perform action -> IFTTT Maker or Amazon skills -> Custom wake word processing Engine == Custom wake word processing Engine -> Google Assistant SDK or Amazon AVS SDK -> Underlying subsystem that perform action -> Google Assistant SDK or Amazon AVS SDK -> Custom wake word processing Engine
?