Deployment-Configuration Steps

This document provides a step-by-step guide for configuring Knovvu Virtual Translator in a demo environment integrated with Genesys. It outlines the required installations, tenant setups, and platform-specific configurations. The guide is intended for application support and presales teams for deploying and configuring the Virtual Translator solution.

Before starting

Before continuing to next steps be sure that you have the information below:

How many agents will use VT, 1 or many? In PoCs or for test purposes, customer might prefer to start using VT with 1 agent only. The configuration steps will change according single or multiple users.
Which languages are required for the project? Language information is required for configuration. If multiple language support is required; will Language Identifier or DTMF (press 1 for English, 2 for German etc.) be used?
Which TTS provider will be used? Sestek TTS or ElevenLabs? -> If ElevebLabs will be used, VoiceID should be delivered by Presales.

1. Hummingbird Installation

Create a ticket to Devops Support team for the deployment Hummingbird to the environment to be used.

You can check the guide on how to create a ticket requesting Hummingbird deployment.

2. Genesys-VT User Implementation Service Installation

Genesys-VT user implementation service is responsible for creating users from Genesys in the CA with same IDs. Every 12 hours, the service automatically gets users from Genesys and creates them in CA, by also assigning their Genesys ID as external ID in CA for user mapping.

Skipping this Step

This step can be skipped if VT will be used for 1 agent only, for PoCs or demo purposes as an example. The users can manually be created in CA and Genesys ID can be manually added as CA external ID.

Create a ticket for the DevOps Support team to install the Genesys-VT User Implementation Service in the target environment.

Example ticket:

xx müşteri/partneri tarafından kullanılacak Genesys-VT entegrasyonunda user implementasyon işlemi yapan genesys-vt-user-integration servisinin VT demolarında kullanılabilmesi için CA'nın xx ortamına aşağıdaki variant adı ile kurulması gerekmektedir. Bu servisin CA'nın güncel imajı ile xx ortamına kurulması konusunda destek olabilir misiniz?

varyant adı: tenantname

3. Creating CA Tenant

Create the CA Tenant in the system. Fill in the cloud request form to request a new tenant from Sales Operations. Select the following features as below, other features can be selected as "No".

4. CA Tenant Configurations

Open the CA tenant, go to Administration > Settings and choose the Virtual Translator service as "Knovvu Core".

Under Administration > Identity Management > Users check the user settings after the deployment of Genesys-VT integration service.

For agent mapping:

If single user is using VT, then Pbx agent ID should be same as the Genesys extension number of the agent. If multiple users are using VT, the external IDs should be same as Genesys agent ID of the users.

If you skipped Genesys-VT User Implementation Service step:

The external ID of the user should be same as the user's Genesys agentID. This will automatically be done if Genesys-VT user implementation service is being used. If not, you must enter the user's Genesys ID as external ID by following the steps below:

In Genesys, go to Admin > People.
Click on the user that will use VT.
Genesys agent ID of the user can be seen in the URL as below:
Go to CA, Administration > Identity Management > Users, enter the ID as external ID of the related user.

5. Creating the Core Tenant

Create the Core Tenant in the system. This tenant will host the project that use the VT capabilities. Fill in the cloud request form to request a new tenant from Sales Operations.

Product must be selected as SR (Cloud API).
ADD TTS or SR should be Yes.
AI should be No.

6. Core Project Configurations

Use the attached VT project and import it into the Core tenant. Then, open the target project and configure it to use the Virtual Translator. This will provide a foundational structure with essential nodes; however, some adjustments are still required for the project to function properly. These include the following:

Creating the credentials

Go to Data Flow > Credentials Page and create the credentials for SR, TTS, and CA applications.

For SR and TTS credentials:

Enter Name, LDM Api ClientId and LDM Api Client Secret in the following format LDMSubscriptionName.ApiClient.xxxx generated while creating this tenant in LDM. If you do not know tenant's LDM information, you may contact salesoperation@sestek.com.
Token Service Address should be https://identity.ldm.knovvu.com/connect/token
Grant Type should be client_credentials
Scope should be Ldm_integration

For CA credentials:

Enter tenant name, clientID and client secret used to access Knovvu Analytics' APIs in the following format CA.TenantName.Analyzer. If you do not know tenant's CA credentials, you may contact salesoperation@sestek.com.
Token service address must be identity server URL with /connect/token added, according to the knovvu Analytics environment used e.g. https://identity.ca.eu.knovvu.com/connect/token
Grant Type should be client_credentials
Scope should be CAExternal

Warning

For the CA credential, don't forget to use the Analyzer role, as below:

Configuring the project

Language Identifier:
This node is essential when the project includes multiple languages and you want to detect the customer's spoken language. Specify the confidence threshold, the list of possible languages to detect, and the minimum audio segment duration.
We recommend setting the confidence threshold to around 0.4 and the minimum audio segment duration to around 500 milliseconds.
SR Nodes:
Enter the credential ID, default language and models to be used within the SR nodes. Make sure to configure it both for the customer and agent channels.
Translation Nodes:
Enter the source language, target language, and API key into the node. The source language represents the callee's language, while the target language corresponds to the other party. Ensure that translation nodes are properly configured for both the customer and agent channels.
TTS Nodes:
For ElevenLabs provider, enter the VoiceId and ModelId information to be used within the project. Make sure to enter a valid API Key for the both agent and customer channels.

For Sestek provider, enter the credential ID, default language and voices to be used within the TTS nodes. Make sure to configure it both for the customer and agent channels.

Sentiment Node:
Enter the language code of the customer to the node.

Important

If an unsupported language is identified in Sentiment node, the project won't run.

CA Realtime Node:
Enter the CA address, credential Id and language code of the analytics project.
RTN Connector Node:
Enter the copilot address and Data Flow external address to the node.
Event Name Changer
This is a custom node required for the VT project to function properly. Since custom nodes are not included in the project's export-import process, it must be created from scratch, connected appropriately, and configured with the following script.

var TestNode = function ()
{
}

TestNode.prototype.onEvent = function (eventInformation)
{	
	//You may create and send a new event to your node's event listeners using raiseEvent(...) method


	if(eventInformation.EventName == "Translation")
	{
		eventInformation.Data.Text = eventInformation.Data.Text; 
	}

	var eventInformation = new EventInformation(
		eventInformation.ChannelIndex, 
		eventInformation.ServiceName, 
		"SR Milestone", 
		eventInformation.Data, 
		eventInformation.StartSampleIndex, 
		eventInformation.EndSampleIndex,
		eventInformation.GroupId,
		eventInformation.Tags
	);
	raiseEvent(eventInformation);
}


createNode = function (nodeConfig, startMessage)
{
	return new TestNode();
}

7. Genesys Configurations

Please refer to Genesys Integration document for the detailed steps.

8. Hummingbird Configurations

The following information should be provided to Bayram Boyraz to complete the necessary configurations. When submitting the request, be sure to specify whether a queue will be used (if there are multiple users using VT) or if the transfer should be made to a specific agent in Genesys. If a specific agent transfer is required, also include the agent’s extension number. If DTMF will be used for language asignment, the information of which number will be used for which language should also be sent.

Core base url
Core Tenant id
Core Client id
Core Client secret
Grant type
Scope
Data Flow wss address
Project name
Agent extension number (optional)
DTMF information (optional)