Deploy AI WhatsApp Agent for handling inbound and outbound conversations
In this guide, we’ll walk you through deploying an AI WhatsApp Agent for inbound and outbound conversations. The AI agent can be used to handle customer interactions on WhatsApp either by responding to incoming messages or initiating an outbound conversation using a pre-approved template.
Step 1: Set up the Trigger for Inbound WhatsApp Messages
In the agent flow, add the “When a WhatsApp message is received” trigger under Incoming Conversations. This trigger activates the mapped agent when a new WhatsApp message is received on the selected WhatsApp number profile.
Design the agent flow to handle different scenarios based on your use case (e.g., responding to customer queries, handling FAQs, etc.). Afterward, test the agent flow to ensure it works as expected.
Outbound WhatsApp Agents are initiated by an API request trigger or by Send Message node within the agent flow.
Meta (WhatsApp) mandates that the first message sent by a Business must be a template message. Once the customer responds to this template, the AI agent can take over and handle the conversation in free-form within a 24-hour window.
Once the customer responds to the WhatsApp template message, the AI agent can carry on with the conversation freely, utilising the designed flow to handle further interactions.
Inbound WhatsApp: When a customer sends a WhatsApp message to your number, the associated agent flow will automatically be triggered.
Outbound WhatsApp: You can either trigger outbound WhatsApp messages via API integration or initiate them directly in your agent flow. Ensure the first message follows Meta’s template policy before the agent can take over the conversation.
Agent Flow Flexibility: Your agent flow can be designed to dynamically respond based on conditions such as a message received, message type (template or free form), or customer response.