hexabot/docs/user-guide/manage-settings/chatbot-global-fallback.md
2024-09-13 18:13:08 +00:00

1.9 KiB

Chatbot Global Fallback

Hexabot's Global Fallback settings allow you to define a default response that your chatbot will use when it cannot understand or respond to a user's query. This ensures a more consistent and user-friendly experience, even in situations where the chatbot might struggle to interpret the user's input.

1. Access Global Fallback Settings:

  • Navigate to the "Settings" section.
  • Select the "Chatbot" view.

2. Configure the Global Fallback Message:

  • Enter Fallback Message: Type the message you want your chatbot to send when it encounters a fallback situation. This message should be polite, informative, and helpful.
  • Example:
    • "I'm not sure if I understand your request. Could you please rephrase it?"
    • "Sorry but i didn't understand your request. Maybe you can check the menu"

3. (Optional) Configure Fallback Triggers:

  • Some Hexabot blocks configurations might allow you to define specific conditions that trigger the global fallback message.
  • This could include:
    • Number of Consecutive Fallbacks: Only trigger the fallback after a certain number of failed attempts to understand the user's query.
    • Specific Keywords: Trigger the fallback if the user's message contains certain keywords that indicate a misunderstanding.

4. Enable Global Fallback

  • Click on the switch button to enable/disable your global fallback settings.

{% hint style="info" %} Tips:

  • Always aim for a positive tone in your fallback message.
  • Let the user know why the chatbot is unable to understand their query (e.g., "I'm still learning" or "I need more information").
  • Offer users alternative ways to get help, such as contacting a human agent or searching for information on a website.
  • Limit the number of times the global fallback message is triggered to prevent a frustrating user experience. If a user repeatedly sends confusing input, you might consider automatically redirecting them to a human agent. {% endhint %}