Skip to main content
All CollectionsFlows
Off-Topic Detection within Flows
Off-Topic Detection within Flows

Learn how off-topic detection works in flows and how it leverages knowledge base retrieval to respond to unrelated questions during flows

Alvaro Vargas avatar
Written by Alvaro Vargas
Updated over 2 months ago

Off-Topic Detection is a powerful feature that enhances your AI Assistant's ability to handle user queries during flows. This feature allows your AI Assistant to understand when a message is unrelated to the current flow and respond appropriately using knowledge base retrieval. Here's what you need to know about Off-Topic Detection within Flows.

Understanding Off-Topic Detection

Off-Topic Detection works seamlessly within your AI Assistant's flows to:

  • Identify messages that are unrelated to the current flow

  • Leverage the knowledge base to respond to off-topic questions

  • Maintain context and return to the flow when appropriate

This feature is automatically enabled for all flows, ensuring a more natural and fluid conversation experience for your users.

How Off-Topic Detection Works

  1. Context Awareness: The AI Assistant uses the flow's name, description, and configuration to understand the current context.

  2. Message Analysis: Each user message is analyzed to determine if it's relevant to the ongoing flow.

  3. Knowledge Base Retrieval: If a message is identified as off-topic, the AI Assistant uses the knowledge base to formulate a relevant response.

  4. Seamless Return: After addressing the off-topic query, the AI Assistant smoothly guides the conversation back to the original flow.

Benefits of Off-Topic Detection

  • Enhanced User Experience: Users can ask questions or make requests unrelated to the current flow without disrupting the process.

  • Increased Efficiency: The AI Assistant can handle a wider range of queries without human intervention.

  • Contextual Relevance: Responses to off-topic questions are tailored using your knowledge base, ensuring accuracy and consistency.

Example Scenario

Imagine a user is in the middle of a "Track My Order" flow:

  1. AI Assistant: "Can you please provide your order number?"

  2. User: "Actually, what's your return policy?"

  3. AI Assistant: [Detects off-topic question] "Our return policy allows returns within 30 days of purchase for a full refund. Now, to help you track your order, can you please provide your order number?"

In this scenario, the AI Assistant seamlessly answers the off-topic question about the return policy before guiding the conversation back to the original flow.

Important Notes

  • Off-Topic Detection is enabled by default for all flows.

  • There are no additional settings to configure for this feature.

  • The effectiveness of Off-Topic Detection relies on well-structured flows and a comprehensive knowledge base.

By leveraging Off-Topic Detection, your AI Assistant can provide a more flexible and user-friendly experience, handling a wide range of user queries while still effectively executing your defined flows.

Did this answer your question?