Agent AI Chat Integration with ServiceNow

This document provides detailed instructions on integrating the Kore.ai Agent AI widget with ServiceNow Desktop for the Chat channel. The integration supports all features of Agent AI. It also includes an end-of-call summary and Custom Data passing, such as agent name and ID, to the Agent AI widget.

Prerequisites

This section outlines the essential components, credentials, and permissions required for integrating Agent AI Chat with ServiceNow.

• Required Components:
  • ServiceNow Developer Instance
    • Washington DC release or higher
    • Configured to receive desktop chats
  • Agent AI Account
    • Enabled Web/Mobile channel
    • Required bot credentials: Bot ID, Client ID, Client Secret, and Widget URL (Credential Location: Go to Flows and Channels > Channels > Digital > Web/Mobile Client > JWT App Details.)
• Roles and Permissions (Optional): Only required for agent transfers from Kore XO Platform to ServiceNow. For more information, refer to Configuring the ServiceNow Agent – Utah and Higher versions.

Interaction between ServiceNow and Kore.ai Agent AI

The following architecture diagram shows the interaction between ServiceNow and Agent AI:

Set Up Kore Agent AI in ServiceNow Sandbox

Setting up Agent AI chat with Servicenow consists of the following steps:

1. Set Up Kore Agent AI in ServiceNow Sandbox
   • Download from the ServiceNow Store (recommended) Or
   • Install Using an Update Set
2. Set Up Agent AI App Configuration
3. Test the Configuration
4. Kore XO Configuration (Optional)
5. Using the Agent AI widget

Step 1: Download from the ServiceNow Store (recommended)

1. Get the Agent AI by Kore.ai application from the ServiceNow store.
2. Click the GET button to proceed with the entitlement process of the application.
3. Select the organization name where you want to install this application.
   

or

Install through an Update Set

1. Ask the Kore representative for the relevant Update Set XML file.
2. Search in All > filter navigator as “Retrieved Update Sets,” which is under “System Update Sets”, and open it.

3. Under Related Links, click Import Update Set from XML.

4. Open the update set Agent AI by kore.ai (This is the official name for Kore Agent AI).
   

Preview Update Set

To preview the update set, click the Preview Update Set tab.

Commit Update Set

This option is enabled after the successful preview. You must click Commit Update Set.

Tips

If you get any error (refer to the following screenshot) in the preview step, follow the below steps to resolve the error.

Steps to resolve the error:

• Close the Update Set Preview dialog window.
• Go to Update Set Review Problems, and select all.

• On the right-side, select Accept remote update from the Action on selected rows… dropdown list.

  Note

  Multi-language support for Agent AI is available only with the Update Set.

Step 2: Set up Agent AI App Configuration

Application Name: Agent AI by kore.ai

This section details the steps to set up the Agent AI app configuration.

Add Agent AI Bot configuration in the ServiceNow Custom Table

1. Change your ServiceNow scope from Global to Agent AI by kore.ai.

2. Go to All > Filter navigator, and search Kore_configuration> kore-config-customtable.

   

3. Click New.

4. Add the Agent AI URL, Bot Id, Client Id, Client Secret (check Prerequisite, and Table Key (give the Table Key as koreai).

   Note

   Client Secret & Token fields are masked for security reasons.

5. Language Code: For any language other than English, select the language code from the dropdown list. The default language code is English (En).

6. Click Submit.

   Note

   The AAtoken field must be empty for the widget to load. The Token field automatically populates at the runtime.

   

Step 3: Test the Integration

• Agent: Sign in to the Servicenow Agent Console/Workspace(Eg- Service Operation Workspace) with your agent credentials (agent-specific permissions) and be available to receive chat on the SNOW desktop.
  • Minimum permissions/roles required for an agent are Itil and awa_agent.

• Customer: To initiate a chat from the customer side, sign in to ESC (Enterprise Service Center) of your Servicenow instance and initiate a chat.

  Note

  ESC portal link: <Servicenow domain>/esc.

  

• Agent: Once the agent accepts the incoming chat, an Interaction Record page opens, and the Agent AI by Kore is the first icon on the contextual side panel.

  • The agent gets a notification of the incoming chat.
    

  • Once the agent accepts the chat, the Agent AI widget appears on the right contextual side panel.

    

  • Once a conversation ends, the conversation summary is generated in the Agent AI widget.

    

Step 4: Kore XO Configuration (Optional)

This step is required if the chat client is routed through the Kore XO bot builder, and post agent transfer, it lands into the Servicenow Agent workspace. For more information, refer to the Configuring the ServiceNow Agent – Utah and Higher versions.

For passing the language code dynamically from Kore XO to the Agent AI widget inside Servicenow, add the following javascript code inside a script node of Kore XO dialog task before the Agent Transfer node of Servicenow. Without this script node, the language code will not be automatically sent to the Agent AI widget.

let metaData = {
    "payloadFields" : {
            "langCode":context.currentLanguage
    },
    "headerFields" : {
            "token" : ""
    }
}
agentUtils.setMetaInfo("ServiceNowMetaData", JSON.stringify(metaData));  

Note

If a Language Code is sent from Kore XO to Servicenow, then the Agent AI widget will load according to that language code. So, the XO language code will always take precedence over the language code set inside the Servicenow custom table in Step 2.

Step 5: Using the Agent AI widget

From the integration perspective, along with all the features and capabilities of Agent AI Introduction to Agent AI, agents on Servicenow have the flexibility to use the following additional features:

• Send / Copy Buttons: Agents can use the Send and Copy buttons on the Agent AI UI to directly send and copy data from the Agent AI widget to the customer.
  

• Conversation Summary / End Of Conversation: The conversation summary is displayed in the Summary box on the Agent AI widget after the end of the conversation when the conversation is closed/ended by agents or customers. Agents have the flexibility to copy this summary on their notepad or save it. If the Submit Summary option is used, the summary is saved inside the interaction table of that conversation.
  

• Conversation logs: Agents or supervisors can check the chat transcript along with the Agent AI summary on the Interaction record page.