Chatbots are great, but let's face it, they sometimes need a human's help to deal with a complex query. To address this need we built a Zendesk live chat integration that lets your chatbots seamlessly handoff conversations to an agent when a user requests it. In this article we'll cover how you can use this integration.

Before we begin, it is worth noting that to complete this integration, you will need:

  1. A Zendesk Enterprise Account
  2. A Tars Chatbot

If you have both of these things in place, we can begin the integration process:

https://www.loom.com/share/f8379ff79bbb473b8a2d8eff23eb622f

Step 1: (Copying account key from Zendesk dashboard)

Login to your Zendesk account

On the right side of the navigation bar you will find a profile icon, select it.

Select the "Check Connection" option and a dialog box like the one in the image below will popup. From this dialog box, copy the "Account Key."

Step 2: (Setting up your chatbot)

In your chatbot, create a gambit, name it "init" and add "Do you want to speak to a Live Agent?" as a message bubble, with Buttons as input and "Yes" and "No" as the button options (like the below image).

Create another gambit, call it "live_agent", and then from the dropdown at the top of the gambit select the "Zendesk Chat" option.

You should see the following config area appear on your screen:

It is important to note, that to begin the live chat, the only compulsory detail that the Zendesk Chat integration needs is the Account Key, you copied in Step 1. Go ahead and paste that "Account Key" you copied from Step 1 in the Account Key field.

Step 3. The Optional Fields

While the other fields (visitor's name, email, phone and start message) are not compulsory, we highly recommend that you collect them as it will help your agent serve the user better.

To input, these fields, first add gambits that collect the necessary details from the user and connect them before the live_agent gambit. You can see how we've set it up below for reference.

Switch back to the live_agent gambit and use data referencing to populate the fields with the appropriate details.

Save the gambit and your agent will now be given all the information they need to serve the visitor.

Please note: The user can end the live chat when he/she types "End Livechat"

Step 4: End the live chat conversation

At this point, your chatbot will seamlessly hand off the conversation to a human agent on Zendesk, so we're done right?

Well, not exactly.

After the agent on Zendesk is done serving the visitor or if the visitor ends the live chat conversation, the conversation will be handed back to the chatbot.

This can be useful if you want the chatbot to handle easy-to-administer, repetitive processes, like leadgen or feedback collection.

To account for this reverse handoff, be sure to create additional gambits after the live_agent gambit to handle the conversation. In the following image for example, once the live chat ends, the user will move on to the "livechat_rating" gambit, where they can rate their experience.

And with that, you know everything that you need to know to integrate your TARS chatbot with Zendesk Live Chat.

Did this answer your question?