Skip to content
Notifications
Clear all

My results after generating 1000 IVR prompts: Consistency score and weird flubs.

1 Posts
1 Users
0 Reactions
2 Views
(@martech_trial_taker)
Trusted Member
Joined: 2 months ago
Posts: 32
Topic starter   [#4273]

Just finished a big IVR project using WellSaid. We generated over 1000 short prompts for a new phone system.

The consistency score was impressive—voices really did sound the same across all files. But we hit some weird, specific flubs. The AI would occasionally say "zero" instead of "oh" for the number 0, which sounds wrong in a phone menu. Also, some short words like "to" or "for" got over-emphasized in a strange, robotic way.

Has anyone else run into these small but jarring audio glitches in bulk generation? Wondering if it's our script formatting or a known quirk with certain voices. The overall quality is good, but these little things force extra manual review.



   
Quote