Understanding and Using Intents in Chatflows

Intents act as specialized entry points in Chatflows, allowing the AI Assistant to trigger specific chatflows based on customer queries.

Luca Micheli
Written by Luca MicheliLast update 2 months ago

What is an Intent?

In the context of Chatflows, an "Intent" serves as a specialized entry point designed to trigger a specific chatflow when a customer's query matches it. Unlike other Chatbot systems where you may need to add If and Else with branches, our AI allows you to attach intents directly to the AI Assistant, making the process more streamlined and focused.

How Does AI Assistant Use Intents?

Our AI Assistant works in real time to interpret customer queries. It has two main functions:

  1. Replying to Queries: The assistant can directly answer customer questions based on pre-defined responses or information from the help center.

  2. Identifying Intents: The assistant can also recognize when a customer's question or statement matches an intent you've defined. Once an intent is identified, the corresponding chatflow is triggered automatically.

As you can see below, there are no branches to create here for the intent. The AI assistant will always recognize the intent if the user mentions it during a conversation.

Setting Up an Intent

If you're looking to cover a new customer intent, the process is simple:

  1. Create a New Chatflow: Start by creating a new chatflow tailored to the intent you want to cover.

  2. Select 'Intent' as Entry Point: Choose 'Intent' from the entry point options dropdown menu. This specializes your chatflow for that intent.

  3. Define the Intent: This involves three steps:

    • Name: Assign an internal name for the intent for easier identification. E.g., "Refund."

    • Description: Add a concise description to help the AI understand this intent. E.g., "The user wants a refund."

    • Examples: Include up to 5 examples of phrases or questions a customer might use to request this intent. This helps the AI understand different ways the same request might be phrased. E.g., "I'd like a refund for last month," or "Can I cancel my subscription and get my money back?"

By following these steps, you create a targeted response system that enhances customer interactions, making your support more efficient and effective.

Now, you can send a tailored message or use powerful, dedicated AI Missions to handle that intent.

Check out AI Missions to unleash the power of complex support automation.

Did this answer your question?