Introduction

This document describes how to configure Webex AI Agent for Contact Center Enterprise (CCE).

Prerequisites

Requirements

Cisco recommends that you have knowledge of these topics:

• Unified Contact Center Enterprise (UCCE) Release 15.0
• Customer Voice Portal (CVP) Release 15.0
• Customer Voice Portal (CVP) Call Studio
• Cisco Virtualized Voice Browser (VVB)

Components Used

The information in this document is based on these software versions:

• UCCE 15.0
• Webex CCE

• CVP 15.0

• Cisco Virtualized Voice Browser (VVB)

• Cloud Connect

The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, ensure that you understand the potential impact of any command.

Webex AI Agent Overview

The Webex AI Agent is a virtual artificial intelligence assistant integrated into Cisco Webex CCE or CCE. It is designed to improve service and support by the used of automated interactions, assist live agents, and deliver insights powered by analytics.

Background

The AI Agent is built to engage with contact center users, interpret the input, generate appropriate responses based on how it is trained, and execute predefined tasks. This guide outlines the steps needed to implement an AI Agent within Contact Center Enterprise (CCE).

The Webex AI Agent is introduced as part of the new CCE 15.0 release.

Webex Agent AI Architecture

Webex Agent AI Architecture

Native: Cisco in-house solution for Virtual Agent

This is the type of Virtual agent that you configure in this lab.

Webex Agent AI Components

Webex Agent AI Components

In both the CCE on-premises solution and the Webex CCE (WxCCE) tenant, several core components form the foundation of the Webex AI Agent solution. These components span both on-prem and cloud environments and work together to enable intelligent, AI-driven interactions.

On-Premises Components

• Cisco Customer Voice Portal (CVP):
  CVP acts as an Interactive Voice Response (IVR) system integrated with CCE/WxCCE. It enables automated self-service and call routing. CVP collaborates with the VVB and Speech Server to integrate the AI Agent within the IVR flow.

• CVP Call Studio: a development environment used to design and build voice applications for CVP

• Cisco Virtualized Voice Browser (VVB) / Speech Server:
  VVB handles multimedia and IVR processing. The Speech Server communicates caller requests to the AI orchestration services using gRPC protocol.

• CCE Core:
  CCE Core includes key components such as Peripheral Gateways, the Router, and Admin Workstation (AW). These work together to process requests and determine the appropriate destination. The AW is used to configure and manage the entire CCE environment.

• Cloud Connect:
  Cloud Connect acts as the secure bridge between the on-prem CCE infrastructure and Cisco cloud services, enabling seamless integration.

Cloud Components

• Webex Cloud Services – Control Hub: a centralized management interface for configuring Webex cloud solutions. From Control Hub, users can provision and launch the AI Agent Studio

• AI Orchestration Services: Cisco’s cloud-based orchestration layer that receives incoming requests from the CCE or WxCCE system and routes them to the Webex Connect AI Agent via a Cisco Connector

• Webex Connect AI Agent: a cloud-hosted AI Agent powered by Webex Connect, designed to handle interactions intelligently

• AI Agent Studio: a visual interface used to design workflows and scripts for the Scripted AI Agent, enabling tailored engagement experiences

Configure Scripted And Autonomous AI Agent on CCE

For information on how to create a Scripted or Autonomous agent, please refer to this document.

Configure Webex Services - Control Hub

Webex Control Hub serves as the central interface for managing your organization, users, service entitlements, and the configuration of Webex Contact Center and other services. It enables provisioning of all Webex services within a single customer organization.

Here are the Webex AI Agent configuration tasks you can test within your Control Hub organization:

• Register Cloud Connect
• Launch the AI Agent Builder

Register Cloud Connect

Step 1. Open a web browser and navigate to Control Hub: admin.webex.com

Control hub signin

Step 2. In Control Hub, navigate to the Contact Center section from the left-hand menu.

Control hub overview

Step 3. On the Contact Center page, click Cloud Connect to begin the registration process.

Control hub cloud connect

Step 4. On the Cloud Connect page, click Add Cloud Connect.

­

Add cloud connect

_­

Step 5. Enter the Name and the Fully Qualified Domain Name (FQDN), then click Register.

add cc cluster

Step 6. In the Allow Access window, check the box for Allow Access to the Cloud Connect Node, then click Continue.

Allow Access to CC

Step 7. Once you see the confirmation that registration is complete, you can close the window.

Cloud connect registeration complete

Configure CCE

The steps required to prepare CCE for Webex AI Agent implementation are:

• Cloud Connect registration
• Cloud Connect proxy configuration
• Cloud Connect Feature Management status

Cloud Connect Registration

Step 1. Open a Web browser and navigate to the PCCE Web Administration Tool also known as the Single Pane of Glass (SPOG).

Step 2. Login with the administrator credentials.

Step 3. In the Web Admin interface, select Infrastructure Settings, then choose Inventory.

Inventory

Step 4. On the Inventory page, locate the Cloud Connect server. Ensure its status is green and no alerts are present.

In this lab, only one Cloud Connect Server (Publisher) is shown. In a production environment, you typically see two servers: Publisher and Subscriber.

CC Publisher

Step 5. Click on the Cloud Connect server to view the Admin credentials and synchronization status with the AW server. Since no changes are needed, click Cancel to exit.

Save

Cloud Connect Proxy Configuration

Step 1. In the SPOG, navigate to the Overview page and scroll down if needed to locate and select the Features card.

Feature

Step 2. On the Features card, click Cloud Connect Integration.

Feature selection

Step 3.  View the registration status of Cloud Connect along with the proxy configuration details.

CC Integration

If your environment requires a proxy to communicate with Webex Services, enter the proxy details used by Cloud Connect.

For example: abc.cisco.com:8080

By default, HTTP uses port 80, but you can specify a different port number in the proxy configuration.

In this lab, a proxy is not required, so it is not configured.

Step 4. Click Overview to return to the main menu.

Overview

Cloud Feature Management Status

Step 1. Log in to the Cloud Connect Command Line Interface (CLI).

Step 2. In the Cloud Connect CLI, enter the mentioned command:

show cloudconnect featureflagmgmt status

You must see that the status is ACTIVE.

Active status

Note: To enable the Feature in Cloud Connect, Information as the SDK key  is required. All the information required is  provided by Contact Center Product Manager

Next, to verify the features enabled for this Webex tenant (including those needed for AI Agent implementation), you can run a specific API call in a web browser.  

Step 3. Open a Web Browser and execute the mentioned API.  

https://finesse1.dcloud.cisco.com/desktop/api/DesktopConfig

API Output

Step 4. In the API response, check the Pretty-print box to format the results for easier reading.  

API Output

In this lab, since you have configured only a scripted AI Agent, the only required feature that the Cisco Product team needed to enable for this tenant is Hybrid_AI_Agent_Scripted, which you see is enabled.

Configure CVP Call Studio Application - Scripted AI Agent

In this example, the Call Studio Application has been developed and it serves as a sample for the implementation in your Call Studio Application.

Step 1. On CVP, click the Call Studio icon on the desktop to launch the application.

CVP

Step 2. In CVP Call Studio, scroll down and expand the Sample application project.

Application

Step 3. Within the Sample Application folder, click on app.callflow to open the call flow.

Step 4. You now see the call flow for this application. It is a simple flow designed to demonstrate AI Agent interaction. The application begins the conversation with the virtual agent using the VAV (Virtual Agent Voice) element, labeled

TrackerBotStart.

After the virtual agent finishes the conversation:

• The application either routes the call to a live agent, or
• Continue the conversation with the virtual agent until it is complete, depending on the result.

Click the Virtual Agent Voice element (TrackerBotStart) and go to the Settings tab to open its configuration.

Step 5. Add the AI agent information. For the Webex AI Agent type, choose Scripted. Once selected the Bot ID field appears.

Copy the Bot ID from the AI Agent Studio and paste it into the Bot ID field.

Copy agent id

This is the id that you need to copy and paste: 68346862640ce715aab84ca7

Call Studio Screenshot

Secure Logging: This setting controls whether potentially sensitive data from the element is logged. If enabled (set to true), the element output—such as query text, fulfillment text, and JSON—from the AI Agent is masked. For now, leave Secure Logging set to false.

You are required to pass an event to the AI Agent. Make sure to set the Event Name to: welcome_event

Call Studio Screenshot

You can send information to the AI Agent by adding variables and values in the Event Data Field. Please enter the mentioned variables and their corresponding values there.

Name: name

Value: your name

Name: place

Value: your location

In this example,

Name: name

Value: Marco

Name: place

Value: India

Call Studio Screenshot

Call Studio Screenshot

Step 5. Now, we review how to gracefully handle the CVP call flow when resources are unavailable or a gRPC error occurs during integration with the AI Agent.

Right-click the VAV element ((TrackerBotStart)) and check the Exit States.

Call Studio Screenshot

By default, the error, VXML Event-noresource and VXML Event-badfetch are part of the element. When any of these errors occur, the application is designed to fall back by playing an audio message and then transferring the call to a live agent.

In this Call Studio application, notice that an audio prompt is automatically played after one of these errors is triggered, ensuring a smooth and informative experience for the caller.

There is no need to add additional events unless you want to handle a specific case. If you do wish to add a new VXML event, click on the VAV element and on the Event tabs click ADD.

Call Studio Screenshot

A new event with the name New_Event1 is added. Select it to modify the name and configure the event settings.

Call Studio Screenshot

Rename it to AIAgentEvent. Leave the Event Type set to VXML Event. In the Event List, you can choose error.noresource.

Call Studio Screenshot

Now, direct the new exit state to the audio element that handles other errors, such as the Error Message element. To do this, right-click the VAV element, select Exit States, choose the new event (AIAgentEvent), and drag its arrow to the Error Message element.

Call Studio Screenshot

At the end you see:

Call Studio Screenshot

Step 6. Now, we review the decision node. Click on the Decision node and choose Use Decision Editor.

Call Studio Screenshot

Step 7. You see options to send the call to a real agent, terminate the session, or use a custom exit. These actions rely on three data elements that depend on the result of the AI Agent. Once reviewed, click OK.

Bot start decision

Step 8. After the interaction with the AI Agent ends, this Call Studio application evaluates the results. Based on that evaluation, the conversation with the AI Agent can continue using a second VAV element.

If you scroll down in the Call Studio application, you find the second VAV element named TrackerBotOrderStatus.

Add the same Bot ID to this element and keep the existing Event Name and Event Data as they are already configured.

This is the ID that you need to copy and paste: 68346862640ce715aab84ca7

Call Studio Screenshot

Keep the Event Data and Event Name unchanged, leaving them as already configured.

Call Studio Screenshot

Step 9. You are now ready to save and deploy the application. Click the save icon to save your changes.

Step 10. Validate the application. To validate the application, right-click on the Sample application and select Validate.

Call Studio Screenshot

Step 11. Deploy the application. To deploy the application on the VXML Server, right-click on the Sample Application and select Deploy.

Call Studio Screenshot

Step 12. In the next window, keep the default settings with the Sample Application checked and the folder set to C:\Cisco\CVP\VXMLServer, then click Finish.

Step 13. On the VXML server update the application. Navigate to C:\Cisco\CVP\VXMLServer\applications\NativeAI_Scripted\admin, then, run updateApp.bat by double-clicking it or executing it in a command prompt.

Step 14. In the new command window that opens, type yes to confirm that you want to update the application.