What's my plan?

Available on all Suite plans

This functionality is part of AI agents.

If you’ve created a conversation bot, there are several standard responses you can customize to define the bot's default responses to customer input during a live conversation

This article contains the following sections:
  • Overview of the standard bot responses
  • Setting up the standard responses

Overview of the standard bot responses

The standard bot responses include:

  • Start of conversation: Initial greeting when a messaging conversation is started.
  • If the bot finds relevant articles: Response when a possible solution is found in your help center.
  • After the bot answers: Response to customer feedback.
  • If the bot needs clarification: Response when the bot can't determine an answer that matches a customer’s comment.
  • If a question matches more than one intent (requires an intent model): Response to a comment that matches more than one intent.
  • If the bot can't answer a question: Response when the customer's comment does not match any answer.

Setting up the standard responses

The bot has several standard responses that you can edit as needed to customize your bot's responses.

To edit the bot standard responses

  1. In Admin Center, click Channels in the sidebar, then select Bots and automations > Bots.
  2. Click Manage bots.
  3. Click the name of the bot you want to edit.
  4. Click the Behavior tab, then click each message to edit it:
    • Start of the conversation
    • If the bot finds relevant articles
    • After the bot answers
    • If the bot needs clarification
    • If a question matches more than one intent
    • If the bot can't answer the question

Each response is saved automatically as you edit it, but updates will not be presented to customers until you publish the bot.

Start of the conversation

When your customers first open a messaging conversation in the Web Widget or mobile SDK, they’re met with an initial welcome message.

There are two types of welcome messages:

  • A standard, configurable welcome message
  • A specific answer in your conversation bot

Configuring the standard greeting message

By default, the bot is configured with a basic greeting message and a suggested answer. You can update the greeting message, add or remove answers (up to 10 answers), or leave them as-is. Answers selected here are presented to your customers as options.

If you opted in to generative AI when you created the bot and you are using bot personas, you can choose to generate variations of the greeting message based on the selected persona.

To configure a welcome message

  1. On the Behavior tab, expand the Start of the conversation section.
  2. Select Send a greeting and suggest answers.
  3. Enter your initial message or leave the default message (Hi there. Got a question? I'm here to help.).
  4. Select Generate variations (optional) to apply a bot persona to the greeting.

    This option appears only if you have opted in to generative AI for bots.

  5. Use the Answers drop-down to select up to 10 answers to present to customers as preconfigured options.

    One suggested answer is included by default ( Talk to a human ).

Using an answer as the welcome message

You can also choose a custom-built answer to start the conversation. When the end user first interacts with the bot, the selected answer is sent. This option can be used to immediately transfer your customer to an agent, for example.

To use an existing answer as your welcome message

  1. On the Behavior tab, expand the Start of the conversation section.
  2. Select Start with an answer.
  3. Use the drop-down to select the answer that will welcome your customers when they initiate a conversation.

If the bot finds relevant articles

If you have an active help center, you can configure how your bot replies when it locates one or more help center articles that can address the end user's question.

If you have opted in to generative AI for bots, your bot can use generative AI to evaluate articles in your help center, then use that knowledge to provide concise answers within the ongoing conversation. Using this option can get your bot up and running immediately without needing to build answer flows.

If you have a restricted help center, the bot respects the article view permissions set in Guide:

  • If the user is authenticated via messaging authentication, the bot can suggest relevant restricted articles.
  • If the user is unauthenticated, the bot will suggest only public articles.

To set up bot replies for help center results

  1. On the Behavior tab, expand the If the bot finds relevant articles section.
  2. Select one of the following response options:
    • Generate a reply: The bot creates AI-generated answers in response to customer questions based on your help center content. See Using AI to generate replies in a conversation bot. Choosing this type of reply deactivates the "If the bot needs clarification" response.
    • Recommend articles: The bot checks your help center for content related to customer questions and shares any relevant articles.
    • Don't answer based on articles: The bot does not reference help center content. Instead, it displays a relevant answer, or replies that it can't answer the question using the response configured in "If the bot can't answer the question." Choosing this type of reply hides the "After the bot answers" response.

After the bot answers

Note: This section doesn't appear if you have selected the Don't answer based on articles option in the "If the bot finds relevant articles" response.
When the bot presents an answer based on help center content, it follows up by asking whether suggested content was helpful. The "After the bot answers" response lets you configure the bot response when an end user says the answer was helpful or unhelpful.

To configure the bot's response to customer feedback

  1. On the Behavior tab, expand the After the bot answers section.
  2. Click the Helpful answer tab and configure the following options:
    • Message: Enter a message to display when the end user says the suggested article was helpful, or use the default message provided. (Great, knowledge is power. You can ask me another question at any time.)
    • Generate variations (optional): Select this option to apply a bot persona to the message. This option appears only if you have opted in to generative AI for bots.
    • Answers: Use the drop-down to select up to 10 answers to present to customers as preconfigured options.
  3. Click the Unhelpful answer tab and configure the options as described above.

    Offering an answer that includes the Transfer to agent step is a recommended best practice.

If the bot needs clarification

You can set up an automatic response to an end user to ask for more information. This response is sent during a conversation if the bot does not understand the end user's comment. This response is optional. You can activate or deactivate the response as needed.

To request clarification

  1. On the Behavior tab, expand the If the bot needs clarification section.
  2. Select Ask customer to provide more info or rephrase question if you want to turn on this response.
  3. Enter your message or use the default message provided ( I'm sorry, I didn't get that. Could you tell me more or rephrase your question, please?).
  4. Select Generate variations to apply a bot persona to the message.

    This option appears only if you have opted in to generative AI for bots.

If a question matches more than one intent

Note: This bot response is available only if your account has a matched intent model.
By default, the bot will present up to three intents it determines to be relevant, if an end user's comment matches multiple intents. If you turn off the option for the bot to suggest intents, the bot will automatically use the intent with the highest confidence score as the best match to the comment.

When choosing a match, the bot looks at all possible intents, including those that are not associated with an answer. If a matched intent that is not associated with an answer is frequently matched to end users' comments, you might want to assign unmatched intents to answers.

After the initial customizable message, the matching intents are presented to the end user as preconfigured options.

To configure matching intent suggestions

  1. On the Behavior tab, expand the If a question matches more than one intent section.
  2. Select Suggest matching intents to recommend up to three possible intents for the end user to choose from.

    Deselect this option to automatically match the end user's comment to an intent.

  3. Enter your message or use the default message. (Is this what you meant?)

If the bot can't answer the question

This response, also known as the fallback response, is triggered when:

  • There is no answer that matches the end user's question or comment.
  • An intent selected by the end user, as part of the "If a question matches more than one intent" response, does not have an associated answer.
You can configure this response to have the bot recommend relevant help center topics. This section is required. It must be configured before you can publish your bot.

To configure the fallback response

  1. On the Behavior tab, expand the If the bot can't understand a question section.
  2. Enter a message the bot will send if no help center articles are found, or if you do not have an active help center.

    The default fallback response is, Sorry, I can't answer that. Here are some topics that might help though..

  3. Select Generate variations to apply a bot personato the message.

    This option appears only if you have opted in to generative AI for bots.

  4. If you want, use the Answers drop-down to select up to 10 of your already-created answers to present to the customer.
Powered by Zendesk