I have an AI bot I've set up to gather information for creating an account in testing. In English, my name is Stephen, the bot confirms it back to me as S-T-E-V-E-N and I can tell it "no, it's S-T-E-P-H-E-N" or "It's a P-H, not a V" and the AI does an impressively good job of correcting my input.
The bot also has Spanish as a secondary language, and my native Spanish speaker testers are having a miserable experience trying to tell it letters if they need to correct a spelling. Spelling TONY *in the Spanish flow* (when language = es-us) with the letters pronounced "tee, oh, en, why" (as English) works 100% of the time; saying them as "teh, oh, eh-neh, ee-gree-eh-gah" (as a native Spanish speaker would) works less than half the time and is more likely to show input in the execution history as "ponigriaga" than be correct.
I have it set to use the default Spanish Speech to Text (Genesys Enhanced v2). Is there some other setting I should use, or is there some other place I need to configure this? Is this a problem others have had? The TTS is working fine for predefined messages in Spanish.
#Architect------------------------------
Steve Sukanek
Lead Analyst
------------------------------