This document provides a detailed explanation of how to effectively use the Customer Intent node within the Sigma Mind AI platform.

​
Purpose of the Customer Intent Node

The Customer Intent node is a crucial component typically used into AI agents directly following a Send Response node or a Wait node. Its primary function is to intelligently analyze and interpret the incoming messages or replies from customers, enabling the AI agent to understand their underlying intent.

​
Asynchronous vs. Synchronous AI Agents and Node Placement

The placement of the Customer Intent node is dependent on the nature of your AI agent’s communication style:

  • Asynchronous AI Agents (e.g., Email): For scenarios where customer messages are not real-time, such as email-based interactions, it is recommended to place a Wait node immediately before the Customer Intent node. This configuration allows for specific verifications or processes to be completed after each customer message is received, ensuring a thorough analysis before proceeding.

  • Synchronous/Real-time AI Agents (e.g., Chat or Voice): In real-time communication environments like chat or voice interfaces, the Customer Intent node should be positioned directly after a Send Response node. This enables the AI agent to promptly analyze the customer’s response immediately after sending a message, facilitating a dynamic and responsive conversation flow.

​
Core Functionality of the Customer Intent Node

The Customer Intent node operates with an inherent capability to “Wait for new message”. When this node becomes active in the conversation flow:

  • The AI agent enters a passive state, effectively “sleeping” at the preceding Send Response node (or before the Customer Intent node itself).
  • It then actively awaits the arrival of a new message or reply from the customer.
  • Upon receiving a message, the node initiates its analysis process to accurately determine the customer’s intent based on the content of their communication.

​
Configuring Customer Intents

The power of the Customer Intent node lies in its ability to be configured with multiple conditions, allowing it to anticipate and categorize various types of customer replies.

  • Adding Conditions: You can define a range of specific conditions within the node. Each condition represents a potential customer intent.
  • Examples of Intent Definitions:
    • “Customer is agreeing to a discount”
    • “Customer is still insisting for order cancel”
  • Designing Responses: By meticulously defining these conditions, you can program the AI agent to send highly specific and appropriate responses back to the customer, or to trigger subsequent actions within the workflow, based on the identified intent.