AI Agent Panel

Overview

The AI Agent panel connects a caller to a configured AI agent within your Logic+ flow, enabling fully conversational self-service interactions powered by a large language model (LLM). When a caller reaches this panel, they are handed off to the AI agent, which engages them in natural conversation and determines the outcome, either resolving the interaction on its own or escalating to a live agent based on the conversation.

Before You Begin

Before using this panel, an AI agent must already be configured and available in your environment. The agent will appear as a selectable option in the Extension dropdown. Contact your Account Manager or a Care engineer if you need assistance setting up or modifying an AI agent.

Example Use Case

A caller contacts your IVR with a billing question. Rather than navigating a touchtone menu, they are connected to the AI Agent panel, which hands them off to a conversational AI agent. The AI agent gathers details, answers the question, and determines that the issue is fully resolved, routing the caller through the SelfServiced exit path without ever involving a live agent. If the caller's question is more complex or they request to speak with someone, the AI agent routes through the Escalated path instead.

Properties

• Panel Name

  • Enter a descriptive, consistent name for the panel.

• Panel Description

  • Provide a clear description of the panel's purpose and behavior. This serves as inline documentation to help team members understand the panel's role within the flow.

• Initiates Flow

  • Designates this panel as the starting point of the IVR flow. When selected, this panel will be the first to execute when the flow is triggered.

• Redact Data in Logging

  • Disables data logging for this panel to support data redaction requirements. This feature is available upon request and is subject to change.

Fields

• Extension

  • Select the AI agent this panel should connect callers to. The dropdown displays all AI agents currently configured and available in your environment.

Exits

The outcome of an AI Agent interaction is determined by the AI agent itself based on the conversation. The following exit paths are available:

• SelfServiced

  • The path taken when the AI agent determines that the caller's need has been fully resolved without requiring a live agent.

• Escalated

  • The path taken when the AI agent determines that the caller should be transferred to a live agent, either because the caller requested it or the issue requires human assistance.

• Fallback

  • The path taken when an error occurs with the AI agent's underlying language model, preventing the interaction from completing normally. Ensure this path leads to a graceful recovery, such as a message to the caller and a transfer to a live agent queue.

Supported Language Models

The AI agent uses one of the following large language models to power its conversations. The available model is configured on the AI agent itself, not within this panel. If the model needs to be changed or updated, please reach out to your Account Manager or a Care engineer.

OpenAI

Anthropic Claude

Google Gemini

xAI / Alibaba / Other