Prod voice AI: latency budgets, vendor churn, and “10¢ vs $10” routing – quick poll

[joEdufi]

Not sales – quick dev poll to decide if I should invest more in infra.

1. Do you hold a latency budget (e.g., P95 < 400 ms) for voice UX?
2. Do you switch vendors when new models drop, or avoid it due to API/auth/QA cost?
3. Does routing simple queries to cheap/fast models and complex ones to premium resonate?

Short write-up + 30-sec survey (no email required): https://revellovoice.com
Happy to summarize results here if links aren’t welcome.