Use a voice processing prompt to fill any silence that the participant hears while waiting for the large language model (LLM) to generate the bot prompts. These prompts provide audio during long periods of silence in the conversation and are similar to the typing indicators that participants see in digital bots.

Configure the default voice processing prompt

1. In the Bot Flow Settings view, click the arrow next to Voice Processing.
2. In the Voice Processing Prompt field, select the Literal expression mode and then select one of the available system or user prompts. For more information about prompts in Architect, see Prompts overview.
3. (Optional) Click Add new prompt if you want to add a new prompt to the list. For more information about prompt creation, see Configure prompt resources.
4. In the Initial Playback Delay field, set the delay duration until which the bot waits before playing the prompt, in seconds or milliseconds depending on the selected time unit.
5. In the Minimum Playback Time field, set the minimum length of time that the bot must play the voice prompt, in seconds or milliseconds depending on the selected time unit.

Click the image to enlarge.

Click the image to enlarge.

Click the image to enlarge.