Interesting!
By async voice, do you mean you record a voice memo and have that processed into a brief? That definitely sounds like a low-friction mode for quick, offline capture.
The live on-call model is interesting to me because it enables a back-and-forth conversation. The agent asks clarifications, suggests alternate approaches, etc. Something about being able to talk ideas back-and-forth helps me refine my idea more.
But async voice memos feel like a really useful complementary mode...