Use Conversation Settings to configure how your AI Agent interacts with users throughout a conversation. You can customize voice behavior, enable customer memory, define timeout handling, and control message-processing settings to create a more personalized and efficient conversational experience.

From this section, you can:

• Configure Voice Settings for voice-enabled AI Agents, including speech profiles, interruption handling, speech processing, and hold behavior.

• Enable Contextual Memory so the AI Agent can use relevant customer information from previous interactions.

• Configure Timeout Settings to define how the AI Agent responds when users do not reply within a specified time.

• Manage

  Message Controls such as response streaming, filler messages, multiple-user-message handling, and conversation reset behavior.

These settings help tailor the AI Agent experience to your organization's requirements and deliver more relevant customer interactions.

Voice Settings

For voice-enabled AI Agents, you can configure the following settings:

Speech Profile

Select the speech profile used by the Voice AI Agent. A speech profile is required for all voice-enabled AI Agents.

First Message

Configure the message played when a voice conversation begins. Use this option for simple greeting or opening messages. For advanced conversation-starting logic, use a Qualifying Task instead.

For each voice reply, you can configure:

• Enable Conditions to control when the response is played.
• Voice Reply to define the text converted to speech.
• Say As to control the pronunciation of values such as dates, numbers, currencies, and phone numbers.
• Modulation to adjust speech characteristics such as pitch and speaking rate.
• Audio Variable to play dynamic audio content.
• Record Audio to record a custom voice response.
• Play Preview to preview the configured response.

• Add Content (+) to add variables or additional audio content.

Barge-in Functionality

Control whether users can interrupt a voice response while it is playing. When enabled, configure:

• Sensitivity: The minimum speech-completion percentage required before user speech is detected.
• Time Prompt: The minimum time the AI Agent must speak before interruptions are allowed.

• Barge-In Response Time: The time the system waits before processing an interruption after user speech is detected.

The highest applicable value is used to determine when interruptions are permitted.

Speech End Threshold

Specify the duration of silence required after a user finishes speaking before the AI Agent processes the input.

Global Hold Settings

Enable hold handling for voice conversations. When a configured hold intent is detected, the AI Agent automatically executes the hold workflow.

Configure:

• Hold Intent to identify the intent that places the conversation on hold.
• Hold Timeout to define how long the AI Agent waits for a user response while on hold.
• Hold Response Message to acknowledge that the conversation has been placed on hold.
• Conditional Responses to display hold messages based on defined criteria.
• Barge-in Functionality to allow users to interrupt hold messages.
• Voice Reply to define the hold message played to users.

• Add Voice Reply to create multiple hold-response variations.

Default Values

Unless modified, the default Voice Settings are:

• Sensitivity: 30%
• Time Prompt: 500 ms
• Barge-In Response Time: 800 ms
• Speech End Threshold: 400 ms
• Barge-in Functionality: Enabled

• Global Hold Settings: Disabled

Contextual Memory

Use Contextual Memory to enable the AI Agent to retain and use relevant customer information across conversations, allowing for more personalized and context-aware interactions.

Timeout Settings

Use Timeout Settings to define how the AI Agent handles situations where users do not respond within the expected timeframe.

• Enable Timeout Settings: Turns timeout handling on or off.
• Wait Time for User Reply: Specifies how long the AI Agent waits before a timeout occurs.
• Timeout Response: Configures one or more responses sent after a timeout.
• Add Voice Reply: Creates multiple timeout responses with different conditions.
• Repeat Count: Defines how many times the timeout response is repeated before the fallback action runs.
• Timeout Task: Specifies the fallback task executed when all timeout attempts are exhausted. 

Message Controls

Use Message Controls to manage additional messaging behavior for your AI Agent.

• Stream Response: Streams AI-generated responses in real time instead of waiting for the full response to be generated.
  

Note: For RAG workflows, streamed responses are displayed token-by-token. When streaming is enabled, output guardrails and post-processing are bypassed, and the AI Agent Response variable does not need to be configured.

• Filler Message: Displays a temporary message while the AI Agent processes a request.
• Multiple Fan Messages: Controls whether users can send multiple messages while the AI Agent is processing a previous request. 
• Reset Context Keywords: Defines keywords that automatically reset the current conversation context and start a new conversation. Common examples include Hi and Reset.