Localizing Virtual Agent conversations

  • Release version: Yokohama
  • Updated January 30, 2025
  • 2 minutes to read
    • Summarize
    Summarized using AI
    This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete.

    Summary of Localizing Virtual Agent conversations

    The ServiceNow platform offers comprehensive methods to localize Virtual Agent conversations to support multiple languages effectively. Using the ServiceNow Localization Framework, customers can manage the entire translation workflow, whether leveraging machine translation, third-party translation management systems (TMS), or a combination of both. This framework standardizes the translation process, defines roles and responsibilities, and enforces business rules such as auto-translation, publication, and approval requirements.

    Show full answer Show less

    Key Features

    • Localization Framework: Centralizes and tracks translation workflows, including request fulfillment options like machine translation, TMS, and file export/import.
    • Role Management: Assigns specific roles—Localization Requestor, Localization Fulfiller, and Localization Editor—to streamline responsibilities throughout the translation lifecycle.
    • Support for Topic Discovery Methods:
      • Keyword Topic Discovery: Translates chat conversations and associated keywords.
      • Natural Language Understanding (NLU) Topic Discovery: Requires localization of both topics and NLU model groups, with translation requests handled via Virtual Agent Designer and NLU Workbench.
    • Translation Management: Enables viewing, requesting, editing, and testing translations by model and language within Virtual Agent Designer.
    • Bulk Translation Requests: Supports mass translation submissions for multiple topics across models using the [syscstopic] table list view.
    • Localization Methods in Scripts: Provides mechanisms to author conversations so untranslated content gracefully falls back to original text until translations are available.
    • Language Support: Supports specific languages for entity extraction in NLU models, with intent matching across others; language support varies for IBM Watson Assistant and Microsoft LUIS integrations.

    Prerequisites and Setup

    • Activate the ServiceNow language plugins for all desired languages (the Localization Framework is included by default).
    • Configure preferred translation modes (machine translation, TMS, or email) in the Localization Framework.
    • If using NLU, enable languages in Virtual Agent settings and add secondary languages to NLU model groups as needed.
    • Assign localization roles to team members according to their responsibilities.
    • Avoid embedded quotes within translatable strings, as these are currently unsupported for localization.

    Practical Outcomes for ServiceNow Customers

    By utilizing the Localization Framework and following the outlined processes, customers can efficiently manage multilingual Virtual Agent conversations to enhance user engagement across global audiences. This approach ensures consistent translation quality, streamlined workflows with clear role assignments, and supports both keyword-based and NLU-driven conversational models. Additionally, the ability to handle bulk translation requests and edit translations directly within Virtual Agent Designer simplifies ongoing maintenance and updates. The framework also ensures fallback behavior when translations are missing, protecting conversational integrity.

    The ServiceNow platform provides several methods for localizing Virtual Agent conversations, depending on your needs. Use the Localization Framework to manage all aspects of the translation process, whether you are using machine translation, a third-party provider, or both.

    Use the ServiceNow Localization Framework to standardize and track the translation process for Virtual Agent topics. The framework enables you to define the business process for translating your conversations. Use it to configure the following:
    • How localization requests are fulfilled.

      Options include machine translation, a translation management system (TMS), and exporting and importing files.

    • Who is responsible for each step in the process.

      Virtual Agent uses the following roles:

      • Localization requestor
      • Localization fulfiller
      • Localization editor
    • Business rules for the process:
      • Auto translation and publication
      • Approval required for translation and publication

    To learn more, see Localization Framework. To learn more about localization roles, see Localization roles for Virtual Agent.

    Localizing topics

    The localization process flow depends on the method of topic discovery that you are using for your topics. The following types of topic discovery are supported using the Localization Framework.

    • Keyword topic discovery

      When you use the Localization Framework, the chat conversation and the keywords associated with the topic are translated.

    • Natural Language Understanding (NLU) topic discovery

      If your topics use NLU, both the topic and the model group must be localized. When you request a translation from Virtual Agent Designer, the topic and any optional keywords are translated. You must request translation for model groups from NLU Workbench. For more information about that process, see Translate a multilingual model.

      Once the topic and secondary models are localized, you may need to map the models to the topic before publishing them.

    Prerequisites

    Important:
    Currently, translatable strings such as topic text and bot responses cannot contain embedded quotes. Strings that contain embedded quotes cannot be localized.
    Do the following before you begin managing translations:
    • Activate the ServiceNow plugin for each language you want to support.

      The Localization Framework is also installed by default. For more information, see Activate a language.

    • Configure the translation mode that you want to use in the Localization Framework.

      You can configure machine translation, a translation management system (TMS), or send via email. For more information, see Translation modes.

    • If you are using Natural Language Understanding (NLU), enable languages in Virtual Agent settings.

      For more information, see Enable NLU languages in Virtual Agent settings.

    • If necessary, add secondary languages to your NLU model group. For more information, see Multilingual model management.
    • Assign localization roles to team member groups.

      Users are assigned different roles based on their responsibilities for the translation process. For more information, see Localization roles for Virtual Agent.

    Manage languages by model

    If NLU is enabled for your instance, you can view and manage translations by model and language. Navigate to All > Conversational Interfaces > Virtual Agent > Designer, and then click Manage languages. Choose a model and a language, and then you can do the following:
    • View translation status
    • Request translations
    • View or edit translations
    • Test translations

    Click Manage NLU translations for easy access to the model management page in NLU Workbench.

    For more information, see Localize Virtual Agent topics that use NLU topic discovery.