Dynamics 365 Customer Service with Power Virtual Agents - Part 1 (automation with text chat)

This article is the first part of a Kickstarter of two pieces and part of my "Dynamics 365 automation with bots/voice bots 101" series; in this article and in the second part, I will focus on Power Virtual Agent in combination with Dynamics 365 Customer Service. The first article focuses on a text chat virtual agent with a handover option to a human agent, aka live chat. The second will demonstrate how to voice enable D365 and use voice bots.
Both pieces will demonstrate that it is possible to provide Omnichannel Customer service with human and non-human agents without a classical Contact Center installation.

Motivation

Imagine a situation where your virtual agent is incapable of solving a query of your customer; that's the point where human-to-human interactions immediately come into play.

Configuring the Power Virtual Agent

Modify Escalate topic

First, we need to modify the default "Escalate" topic to display a meaningful message and handover to the human.

Figure 1: standard escalate topic

Modify the "Message" node with a meaningful message and insert a new "Transfer to Agent" node. By now, we leave the handover to the agent node empty, but this is where you could insert a summary for the human agent.

Figure 2: modified escalate topic

Publish the bot

After creating your Power Virtual Agent and modifying your "escalate" topic, it is time to publish your work to make it visible to the outside world.

Figure 3: Publish your bot

OmniChannel Handover

Configure Power Virtual Agent handover

It would be best if you had a working Dynamics 365 Customer Service environment to work with Omnichannel and human handover. As this is a Kickstarter, I will demonstrate how to start from nothing; you could even start with a 30-day free Dynamics 365 Customer Service trial.

First, we need to activate Agent Transfer with Omnichannel.

Figure 4: Configure Omnichannel

Figure 5: Press enable

We enable voice as we will use it in the second part of this Kickstarter.
If there is no active subscription for D365 Customer Service, we can enable a trial and select the created environment to use with Omnichannel.

Figure 6: Start a trial if needed

Figure 7: Enter your email address

Figure 8: Launch trial

Figure 9: Remember your D365 URL

Figure 10: Select your new environment

Create Application Registration

After configuring these initial steps, we need to create an Application Registration in our Azure Active Directory. We go to the Azure Portal and select Application registrations and "+Add".

Figure 11: Select App registrations

We give it a speaking name and select "Accounts in any organizational directory".

Figure 12: Configure App registration

After saving, we can get the Application ID of the Registration (red)

Figure 13: Remember your Application ID

and paste this ID to our Agent Handover Configuration in Power Virtual Agent.

Figure 14: paste in your Application ID

Figure 15: find your bot connected

Setting up OmniChannel for Customer Service

We need to add some extensions to our environment, to add some features, but everything you see below will work without any extension installed.

For only text (messaging) hand-off

Install Omnichannel Power Virtual Agent extension.

Figure 16: Install extensions

If you still see the following warning after installing the Omnichannel Power Virtual Agent extension and don't need voice capabilities, you can safely ignore it.

Figure 17: ignore the warning

For both text and voice hand-off

Install the following extensions in this order:

1. Power Virtual Agents telephony extension

2. Omnichannel Power Virtual Agent extension

3. Omnichannel Voice Power Virtual Agent extension

Configure Omnichannel for Customer Service

Next, we will create a Queue, a Workstream, and Ruleset.

First, go to https://yourtenant.crm.dynamics.com/apps

Figure 18: Go to your D365 Dashboard

Select the "Customer Service admin center App"

Figure 19: Check bots' existence

Create a Queue

We create a separate Queue for Customers asking for human-to-human interaction.

We select the Type "Messaging" as a queue type

Figure 20: Create a queue

Figure 21: see the "users missing" message

and add some agents after completing it.

Figure 22: add at least one user

Create a Workstream

After creating the first queue, we generate a workstream to select the Work distribution mode and the fallback queue.

Figure 23: Create and configure a workstream

Next, we create a chat widget; you could use a different, more customizable chat window configuration. You can find an example in my "Create your first (voice-) bot" article.

Click "Set up chat" and configure it accordingly to the screenshots.

Figure 24: see "live chat required" message

Figure 25: configure live chat - part 1

Figure 26: configure live chat - part 2

Figure 27: configured live chat

Add a RuleSet

Next, we create a ruleset for handling the users.

Figure 28: Create a ruleset

and create a new rule.

We do not need to configure a condition, as we want to allow every user to enter the queue and assume you have a 24/7 service.

Figure 29: Create a rule

Add Bot to Workstream

We add our bot via smart assist bots' settings in the Advanced settings area.

Figure 30: Add the bot to the workstream

Add Context Variables

We need to add a first context variable to the workstream.
We will use the va_Scope Variable.

Figure 31: test the chat to display the context variables

Figure 32: see the context variables

Figure 33: add context variables

Figure 34: Create context variable

Test handover

Let's test the handover.

Go to your workstream, copy the code for the snippet, create an empty HTML-File and paste the copied code into the body.

Figure 35: Copy code snippet

You will end up with something looking like this.

<!DOCTYPE html>
<html>
<head>
<title>Page Title</title>
</head>
<body>
<script id="Microsoft_Omnichannel_LCWidget" src="https://oc-cdn-public-eur.azureedge.net/livechatwidget/scripts/LiveChatBootstrapper.js" data-app-id="your data" data-lcw-version="prod" data-org-id="your data" data-org-url="https://your data.crm4.omnichannelengagementhub.com"></script>

</body>
</html>

Open the HTML File with your Browser and start chatting.

Figure 36: start the dialog

Ask the bot for a human agent.

Figure 37: ask for an agent

In your D365 Agent Dashboard, you should see an incoming chat and can talk to yourself.

Figure 38: answer request

The agent can see the customer's dialog with the bot on the left side.

Figure 39: Explore the interface

And can start to chat with the customer.

Figure 40: customers' view

Conclusion

Self-service helps you provide 24/7 effective and efficient customer service; with the integration of human handover, you can keep your customer satisfaction high.