Situations
      Back to home
      1. Help Center
      2. Pulse
      3. Situations

      Limiting the number of handoff triggers

      This article explains why it is important to limit triggers and how to approach this effectively.

      Setting up triggers for handoff is a crucial part of a well-functioning chatbot. However, adding too many triggers can result in conversations being handed off too quickly and more often than desired. This can lead to inefficiency and a poorer user experience.

      Why limit the number of triggers?

      1. Prevent unnecessary handoffs: The more triggers you set, the higher the chance that a conversation is handed off unintentionally.

      2. Improve user experience: Unnecessary handoffs can frustrate users, especially when their query could have been resolved by the chatbot.

      3. Maintain clarity: A limited number of triggers makes it easier to monitor and manage how the chatbot responds.

      How to limit the number of triggers

      Instead of adding a separate trigger for every situation, use a single general trigger. Combine this with a clear instruction that covers multiple scenarios. This ensures that the chatbot can handle requests more flexibly and efficiently.

      Example instruction for Handoff

      Use the following approach to cover multiple scenarios with a single trigger:

      1. Write a comprehensive instruction: Add a message to Domain Knowledge or Instructions such as:

        "If someone wants to speak to a staff member, file a complaint, request a quote, report an issue, ask about a product, or needs other assistance, let them know you can hand off the conversation to a staff member. Mention that if they really want to speak to a staff member, they should type the words ‘speak to a staff member’ to proceed."

      2. Set a single trigger: Use a general trigger, in this case, “speak to a staff member", to activate the handoff when those words are used.

      Tips for limiting triggers

      1. Use general terms: Choose a single trigger that can cover multiple situations.

      2. Test and optimize: Ensure the chatbot responds correctly in different scenarios and adjust the instructions as needed.

      3. Monitor conversations: Track how often conversations are handed off and refine triggers and instructions based on this data.