I remember some discussion about this early on. Check out these and see if they can help you:
Genesys Cloud CX - Genesys
Also, on the Google side, here are some things to try:
- You can use the built-in events
sys.no-input-1 and sys.no-input-default to trigger different responses or actions when there is no input from the user. For example, you can repeat the question in the same way as the first time for sys.no-input-1, and rephrase the original question for sys.no-input-default. You can also set the transition page to "Current Page" in the event handler to loop the error on the same page. See this answer1 for more details and screenshots.
- You can use a catch node to catch events like "repeat", "silence" and low confidence checks. You can increase a counter by +1 each time there is no input from the user, and repeat the initial phrase/question. On the third try, you can bounce them to an agent. See this answer2 for more information.
- You can follow the voice agent design best practices3 suggested by Google Cloud, which include limiting no-match and no-input repetitions, providing clear instructions and feedback, and using confirmation and disambiguation strategies.
------------------------------
Robert Wakefield-Carl
ttec Digital
Sr. Director - Innovation Architects
Robert.WC@ttecdigital.comhttps://www.ttecDigital.comhttps://RobertWC.Blogspot.com------------------------------