Understanding Answer Bot (for administrators)

Important: Answer Bot is currently only available in English, Spanish, Portuguese, French, German, and Dutch.

Answer Bot uses machine learning to respond to support tickets by sending an automated email that lists potentially relevant knowledge base articles. This encourages self-service in your customer base, and raises awareness of your knowledge base offerings. It analyzes titles, text, and labels to select the best articles in your Help Center.

This article covers the following topics:

• Answer Bot requirements and limitations
• The Answer Bot management page
• Understanding the end-user experience

For information on working with Answer Bot, see the following articles:

• Enabling and setting up Answer Bot
• Working with Answer Bot (for agents)
• Best practices: Six steps to success with your Answer Bot

Answer Bot requirements and limitations

In order to use Answer Bot, you must meet the following requirements:

• Be a subscriber to Guide Legacy or Professional
• Have an active Help Center with at least 10 English-, Spanish-, Portuguese-, French-, Dutch- or German-language articles
• Be set up to receive English-, Spanish-, or Portuguese-, French-, or German-language tickets

Answer Bot has the following limitations:

• It is not available for multiple brands using the hub-and-spoke model; if you're using more than one brands, you must be using Multibrand.
• If you are using Multibrand, Answer Bot will only suggest articles associated with the brand used on the  related ticket. If a ticket is assigned to Brand A, for example, then only articles written for Brand A will be suggested. 

The Answer Bot management page

You can open the Answer Bot management page by clicking the Admin icon () in the sidebar, then selecting Business Rules > Answer Bot.

From here you can view and manage your monthly activity, view and edit Answer Bot triggers, and view recommended triggers for using with Answer Bot.

Note: Answer Bot is not compatible with the {{header}} placeholder, which some customers may still be using in their HTML email template. This placeholder functions with standard email notifications, but produces a malformed header in Answer Bot notifications.

Monthly activity summary

The montly activity summary displays information Answer Bot activity, including:

• Successful resolutions, the number of support requests marked by customers as Solved because of a suggested article.
• Available resolutions, the number of remaining resolutions based on the selected Answer Bot subscription limit.
• Days remaining, the number of days left in the current Answer Bot subscription cycle.

In this section, you can also open the Answer Bot subscription management page, by clicking the Manage button.

Answer Bot triggers

The Answer Bot triggers section lists all active triggers with the Notification: Answer Bot action enabled. It displays the trigger names and usage numbers for the past 30 days.

In this section, you can also open each trigger's edit page, by clicking the Edit button. For more information on creating triggers for Answer Bot, see Creating new triggers for Answer Bot

Recommended triggers

The Recommended triggers section offers suggestions for triggers that run on ticket creation, that could be improved by adding an Answer Bot notification. It displays the trigger name, how often the trigger was used in the previous 30 days, and the estimated number of potential resolutions.

In this section, you can also open each suggested trigger's edit page, by clicking the Edit button. For more information on modifying triggers for Answer Bot, see Converting an existing trigger for Answer Bot.

Understanding the end-user experience

With Answer Bot enabled, the end-user receives an automated email response to their support request. The email looks like a standard auto-response, with a list of suggested articles and other information provided by the placeholders used in the Answer Bot trigger.

For more information on placeholders, see Creating new triggers for Answer Bot.

From here, the end-user can perform a number of actions, including:

• Clicking any of the suggested article links. This opens the Help Center article in a new tab. From there, the user can read the article, click to see to their help request, and indicate whether it helped them answer their question:

  

  • Clicking the request number opens the request in a new tab.
  • Clicking Yes, close my request opens the article in the Help Center, and closes the help request.
  • Clicking No opens an optional feedback window, asking for more information about why the article didn't help.
• Reading the top suggested article. The entire article is included in the email.
• Clicking the Yes, close my request button, for any of the suggested articles. This opens the article in the Help Center, and closes the help request.
• Clicking No beneath the top suggested article opens an optional feedback window, asking for more information about why the article didn't help.

End-users can return to the email and access the links as often as necessary.