Building Guide for Basic Zendesk Bot Template – Help Center

Are you going to integrate your chatbot project with Zendesk, Zendesk Chat, and Zendesk Guide? You can get from zero to hero with our Basic Zendesk Bot Template.

To access the Template, please click here and select the Basic Zendesk Bot Template in the Template overview on top of the Bot List page you will land on (above the fold).

After setting up your bot, it will search your Help Center for relevant articles based upon your end-users questions and offer them additional support either through tickets or chat. Here is an overview of its capabilities:

Use metadata from your Zendesk widget Pre chat form
Present relevant articles from your Help Center
Measure CSAT and offer additional support
Handover to support by ticket or chat
Understanding end users' needs with the help of NLU

To access the Template, please click here.

In this article, we will cover the following:

Different Color Framings
KPI Tag Indications
Selected Module Configurations
Global Connections Configuration

Different Color Framings

On the bot building canvas, you can find several Modules in different color framings. Each color stands for a specific Module category or group. The overview of the color framing can be found on the left sidebar. From here, you can manage the visibility of selected groups.

Color	Module Category	Explanation
	Start	Introduction to the bot conversation followed by a CTA that encourages end-users to ask a question.
	Help Center	Modules that are triggered after the bot performed a Help Center search upon the end-users preliminary question. If any article is found, the bot will display the most relevant and then ask for feedback.
	Fallback	Modules that are triggered when the bot cannot understand the end user's input. Read this article for more explanation: Fallback module
	Fallback ticket	If end-users request ticket support, this flow will create a ticket with their preliminary question and meta-data collected from the pre-chat form. If any comments need to be added, the ticket can easily be updated.
	Fallback chat	If end-users request live support, this flow will check agent availability and perform a handover if possible. Read this article for more explanation: Bot takeover in Zendesk
	Exit	Any Module that relates to the end of a specific conversation

KPI Tag Indications

You can find Key Performance Indicator (KPI) tags assigned to specific Modules.

Webhook KPI tag: Indicates in which Module a Webhook must be implemented in Module Connections.
Start KPI tag: Shows the Starting Module. It is important that the checkbox "Make bot start message" is ticked in this Module's Message tab.
Handover KPI tag: Points to Modules in which a transfer to a real person happens. This transfer can either be connected to live chat or create a message to be received by human agents.
Fallback KPI tag: Marks Fallback Modules triggered whenever the bot does not understand the end user's input.
KPI tag 4: Set for Module categories that contain organization-specific content. You may cluster this content following the Domains you have defined for your bot.
KPI tag 5: Find out how many visits lead to a feedback rating
KPI tag 6: Analyse how many users rated an answer as helpful (positive feedback)
KPI tag 7: Analyse how many users rated an answer as not helpful (negative feedback)

We can now take a closer look at some Module configurations.

Selected Module Configurations

To ease your platform navigation, use Control+F (or Command+F on the Mac) and look for the specific Module titles as listed below.

Module name	Description
Starting Module	The purpose of this Module is to collect your chatbot visitors' meta-data from the pre-chat form, which is set under connections. When using the Zendesk Widget your bot does not start unless it is prompted by the end user to do so - find a thorough explanation in this article. If you do not use the Pre-chat form to collect end-users name and email, we suggest adding modules to your canvas that collect these.
Welcome Module	The chatbot visitor is welcomed and introduced to your chatbot. Expectations are essential to align early, why the fact that short and straightforward visitor messages are preferred is being mentioned.
Help Center search	The chatbot visitor is asked to phrase the question, which will be used for performing a search in your help center. Note that a feature called variation is used here, as the chatbot visitor is likely to come back to this Module. By using the variation feature, you give the bot answers a more human touch. The Modules in the welcome flow are separated so that they can be triggered individually later on. Necessary Webhook: Find the Webhook Template Zendesk Guide Search Knowledge Base on the Webhook Marketplace and insert it between the modules Connection and Fallback. Decide for either illustrating the results as Cards (Zendesk Guide Search Knowledge Base: Show as Cards) or as a List (Zendesk Guide Search Knowledge Base: Show as Text and Links). Follow the instructions on top of the selected Webhook.
Help Center results	If no results are found in your Help Center or end-users report results non-useful, human support is being offered.
Fallback	Besides your welcome flow, end-users can be sent to the Fallback Module if it´s set as the fallback on individual Modules. Note that the Fallback Modules are not showing a dotted line in between. The reason for the missing line is that each Fallback Module is connected with the following one by being part of the previous Module's Fallback: Fallback 2/3 of Fallback 1/3 Fallback 3/3 of Fallback 2/3 Looking at each Fallback Modules' Connections will make this configuration clearer.
Create ticket with new/preliminary question	If support is preferred through ticket one will be created. Passing the preliminary question from your chatbot to the Zendesk environment. If no preliminary question exists yet, it will be captured in the Module Create ticket with new question. Necessary Webhook: Find the Webhook Template Zendesk Support Create Ticket on the Webhook Marketplace and use it in the modules "Create ticket with preliminary question" and "Create ticket with new question". Follow the instructions on top of the selected Webhook.
Update ticket	In case end-user do wish to add comment to the newly created ticket, this can be easily done in this module. Necessary Webhook: Find the Webhook Template Zendesk Support Update Ticket on the Webhook Marketplace and use it in this module. Follow the instructions on top of the selected Webhook.
Check chat availability	For end-users who prefer chat support, a webhook will first check if any agents are available. If true, the bot will perform a handover to your chat department. Necessary Webhook: Find the Webhook Template Zendesk Chat Agent Availability on the Webhook Marketplace and use it in this module before the chat handover. Follow the instructions on top of the selected Webhook.
Agent available	If any agents are available, the end-user will need to confirm the chat before the handover is trigged. Necessary Webhook: Find the Webhook Template Zendesk Chat Handover to Agent on the Webhook Marketplace and use it in this modules first connection. Follow the instructions on top of the selected Webhook.
Bot takeover	When agents ends the chat, you can set the bot to jump back into the conversation and takeover. It´s pretty simple, if you follow these steps: Hit settings on your canvas and click Bot Takeover. Create a connection that "If the ClaimEnd value sent by the API to give back the control to the Bot" exactly match HumanAgentExit go to Bot Takeover.

To check if the Webhooks have been executed correctly, we recommend you insert "Real Fallbacks". Read the article "Real Fallback" for Webhook Failure to learn more.

Global Connections Configuration

If you are in doubt about the general functionality of Global Connections, learn more about it in this article: Basic Insights on Global Connections

Note that in the Support Bot Template, Module Connections are preferred over Global Connections. This means that Global Connections are trigger if not Module Connection applies to the specific end user's or system (in case of third-party integrations) input.

The Basic Support Bot will understand the following Intents at any point of the conversation if Global Connections are not deactivated:

"Goodbye"
"Hello"
"Speak to a Human"