Glad you saw the talk. Just wanted to add that this approach can be used by Speech Science as well. Our speech scientists have compared grammar performance using a random split, while holding the VUI constant.

As far as VUI, we are also currently running an experiment to observe the effects of variations in opt-in prompt wording (By “opt-in prompt” I mean a prompt that gives the caller the option of working with us or waiting in line for an agent). No data yet.

Cheers,
jon bloom

Thanks for the shout-out! It was good to see you. Hope Nuance is treating you well.

Phillip