AI Chatbot Guide v1
AI Chatbot
Version 1

Custom responses

Copy link

Along with the function calls feature, Sendbird also provides the custom responses feature, where businesses can train their AI chatbots with specific questions and predefined answers, enhancing the control over the bot's output. This leads to more coherent and reliable responses, ensuring a consistent and trustworthy customer experience.

Set up a custom response

Copy link
  1. On Sendbird Dashboard, go to Application > Chat > AI chatbot > Custom responses > and click on Create response +.

  2. Enter the basic information for the function.



A name for the custom response. This name is only used for the dashboard.


A key to filter or identify the custom response by. Only letters, numbers, and underscores can be used.

  1. Train the AI chatbot by providing training questions as shown in the image below. You can add up to 10 questions.

You can also generate and add AI suggested training questions.

  1. Test the AI chatbot by asking different questions and see if the training has been successful.

  1. Predefine a custom response to the training question provided above.


Response type

-Function calls & Messages: Use the pre-set function call in addition to a message.
-Messages only: Only send pre-written messages as a response.

Message to send before function calls begins

The message to send to the user before showing the output from a function call.

Message to send when function calls is unsuccessful

The message to send to the user when a function call fails.

Suggested replies

One-click responses designed for fast and easy user engagement.

Incorporate it into your AI chatbot

Copy link

Using the custom responses you created above, you can directly incorporate it into your AI chatbot.

  1. On Sendbird Dashboard, go to Application > Chat > AI chatbot > Manage bots and either click on an existing bot or create a new bot. For more information on how to create a new bot, see this page.

  2. In the bot detail page, you can select and add a response under Custom responses as shown in the image below.