Overview of the lab:

In this Lab, we will go through the tasks that are required to setup Contact Center AI with the Voice Bot and Text-to-Speech (TTS) capabilities.

Table of Contents

Introduction

Lab Objective

  • This lab is designed to help you configure a Google DialogFlow Agent using CCAI (Contact Center AI) on Webex Contact Center and utilize TTS (Text-to-Speech) capabilities.

  • At the end of this lab, you should have a fully functioning bot front-ending the Webex Contact Center and text to speech prompts.

Prerequisites

  • You must have a Google Account created.

  • A credit/debit card (American Express, Mastercard or Visa) is needed to create the Google Billing account.

    Note: No card will be automatically billed any cost, but a billing account needs to be setup to be able to utilize TTS.

  • Lab 2 (IVR Contact Routing Lab) should be completed, as same call flows will be used and expanded upon.

Control hub: https://admin.webex.com

Portal: https://portal.wxcc-us1.cisco.com/portal

Agent Desktop: https://desktop.wxcc-us1.cisco.com

Google Cloud Console: https://console.cloud.google.com

Dialogflow: https://dialogflow.cloud.google.com

Lab Section

1. Setup the Google Account

  • Login to Google Cloud Console with your Google Account credentials.

  • Click on Select a project on top and then on NEW PROJECT on the pop-up window.

  • Enter a name for your project, e.g. TS2021-CCAI and click on CREATE.

  • Make sure you have this project selected.

  • On the navigation bar on the left, click on Billing.

  • Click on LINK A BILLING ACCOUNT and then on CREATE BILLING ACCOUNT.

  • Enter your Account Information (Country, Mobile Phone), accept the Terms of Service and click on CONTINUE.

  • Enter your credit/debit card details and click on START MY FREE TRIAL.

  • Click on CLOSE on the next pop-up window.

  • On the search bar, type text to speech and click on the Cloud Text-to-Speech API.

  • Click on ENABLE.

  • Click on CREATE CREDENTIALS on the top-right.

  • If prompted,choose Cloud Text-to-Speech API from the dropdown menu and check the Application Data option and No, I’m not using them and then click on NEXT.

  • On the Service account details, enter a name for the service account, e.g. TS2021_TTS_SA and then click on CREATE AND CONTINUE.

  • On Grant this service account access to project, search and choose the DialogFlow API Admin role and then click on CONTINUE and then on DONE.

  • Similarly, now on the search bar type dialogflow API and click on the Dialogflow API.

  • Click on ENABLE.

  • Click on CREATE CREDENTIALS on the top-right.

  • If prompted,choose Dialogflow API from the dropdown menu and check the Application Data option and No, I’n not using them and then click on NEXT.

  • On the Service account details, enter a name for the service account, e.g. TS2021_CCAI_SA and then click on CREATE AND CONTINUE.

  • On Grant this service account access to project, search and choose the DialogFlow API Admin role and then click on CONTINUE and then on DONE.

  • To generate the required .json keys, click on the service account you created (e.g. ts2021-ccai-sa@ts2021-ccai.iam.gserviceaccount.com).

  • Click on the KEYS tab and then on ADD KEY > Create new key.

  • Choose the JSON key type and click on CREATE. Make sure you have the key saved locally.

  • Click on Service Accounts on the left and follow the same procedure to download the TTS JSON key as well.

Part 2. Setup Dialogflow Agent & Google Connector

  • Open Control Hub Admin page.

  • Go to Contact Center > Connectors and click Set Up on Google connector.

  • Give a name to the connector, e.g. techsummit_google_tts and click on Upload Authentication Key to upload the .json file key downloaded before. Click on Done.

  • Go to Features tab and click on New template.

  • Click on Virtual Agent.

  • Check only Use For Voice and click on Next.

  • Choose Yes, I have a preconfigured Dialogflow agent.

  • Click on Download Intents and save the .zip file with the two intents locally.

  • Open a new tab and go to Dialogflow page and login with your Google credentials.

  • Click on Create an Agent.

  • On the GOOGLE PROJECT, give a name to the agent, e.g. TS_DF_Agent, choose the project created in the previous step and click on CREATE.

  • On the next page, click on the dots on the top right (next to CREATE INTENT) and choose Upload Intent.

  • Unzip the two intents you downloaded from Control Hub (escalation.json & handled.json) and upload them here.

  • Go back to the Control Hub tab and click on Next.

  • Give a name to your virtual agent, e.g. TS_CCAI_Agent and click on Next.

  • Upload the CCAI/Dialogflow .json key downloaded in the previous part and click on Validate. If all is good, click on Next.

  • Skip the avatar step and click Next.

  • Click on Finish to complete the Virtual Agent creation.

Part 3: TTS, EWT & PIQ

  • Open Flow 3 in the Flow Designer from Lab 2 (IVR Contact Routing Lab).

  • In the Success Play Message block, chose Enable Text-to-Speech under Prompt, choose your created connector (e.g. techsummit_google_tts) and set language as en-GB-Standard-A.

  • Type “Thanks {{Customer_Name}}, we got your information” as your TTS message.

  • Make a new test call and verify that you get a personalized message with the customer’s name based on the PIN provided.

  • For EWT, after the QueueContact1 node, add a Get Queue info block.

  • Under Queue Information, select Static Queue and enter your queue.

  • On success path, add a new Play Message block. Name it EWT_PIQ.

  • Add the TTS connector to it similar to the step above and type “You Estimated Wait Time is {{GetQueueInfo.EWT}} and your position in Queue is {{GetQueueInfo.PIQ}}” as message.

  • Similarly, for the Insufficient Information block, add a new Play Message block and name it PIQ.

  • Again, add the TTS connector and type “Your position in Queue is {{GetQueueInfo.PIQ}}” as message.

  • Save and publish the flow. Test the call flow and make sure you hear the PIQ prompt.

Note: Make sure no agent is available to handle the call to be able to listen to the PIQ prompt.

Part 4: Virtual Agent

  • Create a new call flow in the Flow Designer.

  • Create a flow logic similar to the video guide, i.e. first add a Virtual Agent block after the New Phone Contact.

  • Under Conversational Experience, add the virtual agent you have created (e.g. TS_CCAI_Agent) and enable Make Prompts Interruptible.

  • For Handled output from Virtual Agent, which means that self-service was successful, connect the output to a Play Message block and add a “Call successfully handled by virtual agent!” TTS message to it.

  • Connect that Play Message to a Disconnect Contact block.

  • Respectively, for Escalated output, which means that self-service was not sufficient and we need to escalate this to an agent, connect the output to a Queue Contact block and assign your Queue to that block.

  • Add a Play Music block with some music on hold and loop it to itself to use as MOH while customer is waiting for an available agent.

  • Map that flow to your entry point to be able to make test calls.

  • Make a test call and when virtual agent greets you and asks you what you would want, answer with I don't need any more help. This will trigger the Handled output and terminate the call after a message from the bot.

  • Make another test call. This time, answer with I need help with billing. This will trigger the Escalated output and search for an available agent.

  • Optionally, you can send the transcript of the conversation between the customer and the virtual agent to the agent when connected. To do so, you would need to Enable Conversation Transcipt under Advanced Settings in the Virtual Agent block.

  • Also, you would need to create a CAD Variable to save the TranscriptURL output variable from the Virtual Agent block and pass it to the desktop.

  • Feel free to play around with various phrases as responses to the virtual agent and see if it is able to handle them properly.

Congratulations! Google CCAI Integration lab is now complete!

We would like to keep track of your progress and make sure that we are giving you effective support. Please take approximately one minute to complete the short survey.

Back to top