Document Intelligence for Customer Service

  • Release version: Washingtondc
  • Updated February 1, 2024
  • 3 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 Document Intelligence for Customer Service

    Document Intelligence for Customer Service enables the extraction of pertinent information from emails and case attachments, such as credit card numbers and customer addresses, to enhance case management. Agents can directly review and correct extracted values through the Document Intelligence interface, contributing to continual model improvement through a Human In The Loop (HITL) process.

    Show full answer Show less

    Key Features

    • Predicted Field Values: Fields containing extracted values are marked with an AI icon and labeled as "Predicted from DocIntel." No manual refresh is needed to view updates.
    • Prediction Banner: A banner indicates when predictions are present, not reviewed, or being updated. Once reviewed, the banner is removed.
    • DocIntel Admin Experience: Accessible through the Task Intelligence Admin Console, this feature allows for the creation, configuration, and performance monitoring of document processing use cases.
    • Document Classification: Administrators can create document classification use cases to help the AI recognize various document types, improving document handling efficiency.

    Key Outcomes

    By utilizing Document Intelligence, ServiceNow customers can streamline their case management processes, reduce manual data entry errors, and enhance the accuracy of information extracted from documents. Continuous training and adjustments to the AI model lead to improved performance over time, benefiting overall customer service operations.

    Use the Document Intelligence for Customer Service feature to extract relevant information from email and case attachments, such as credit card numbers or customer addresses, and add that information to cases.

    Figure 1. Document Intelligence for Customer Service extracted fields
    Case details tab displaying Predicted or Predicted from Docintel labels under certain field values, such as Account or Vendor name, alerting agents that the information was extracted by Document Intelligence.

    Agents can review values for extracted fields and make corrections as needed by accessing the Document Intelligence interface directly from the case. From this interface, agents can confirm correct values, fix incorrect values, and continue to train the model. This HITL/Human In the Loop interaction of verifying the recommended values enables agents to refine the model and continually improve performance.

    Predicted field values

    In CSM Configurable Workspace and Core UI, the fields on the Case form that contain Document Intelligence predicted values are identified with an AI icon () and label. These fields also include an information icon that displays a message with additional context about the predicted values.
    Note:
    The system displays the AI icon automatically. The agent does not need to refresh the record page.

    In CSM Configurable Workspace and Core UI, the fields on the Case form that contain Document Intelligence predicted values are identified with the message Predicted from DocIntel.

    Prediction banner

    The system displays a banner at the top of a case record that contains one or more Document Intelligence field predictions. The banner is displayed:
    • When there is at least one field with an auto-filled prediction from a categorization model.
    • When one or more of the extracted fields has not been reviewed by the agent.
    • When the fields are being updated.
    • When the fields cannot be generated or predicted.
    Once an agent reviews the field, the banner is removed.

    The banner is displayed for records in the Case table, extensions of the Case table, and interaction records. The banner is displayed in the Core UI and CSM Configurable Workspace.

    The banner can be enabled or disabled by the sn_csm_ml_task.ui.banner.enabled system property.

    DocIntel Admin experience

    The DocIntel Admin experience provides an easy-to-use interface that you can use to do the following:
    • Create and configure document processing use cases
    • Monitor the performance of Document Intelligence solutions

    The DocIntel Admin experience is available with the Document Intelligence Admin (com.snc.docintel_admin) store app. This app automatically activates the flows and properties required by Document Intelligence for Customer Service.

    Access the DocIntel Admin experience through the Task Intelligence Admin Console:
    1. Navigate to Task Intelligence for Customer Service > Setup.
    2. In the Explore related applications section of the console, select Open DocIntel in the Document Intelligence card.
      Note:
      If Document Intelligence Admin is not installed, the Home page displays a link that you can use to download and activate the application.
    For more information, see the following topics: .

    Use Cases list

    From the DocIntel Admin experience, select Use Cases in the header to display a list of the use cases. The Use Cases list uses a default filter with an OR condition to show use cases for:
    • Application = Task Intelligence for Customer service -or-
    • MLUC ID = MLUC CSM-00003
    Note:
    The user cannot modify these filter conditions from the advanced view.

    Each use case has an MLUC ID number that is automatically assigned when the use case is created. This is an ID number that is used for machine learning usage tracking. The Task Intelligence for Customer Service application manages the MLUC ID for Document Intelligence for Customer Service use cases.

    Document classification

    Users with the system administrator role can create document classification use cases and define the classes or categories for the AI to detect and apply to documents. This is useful in situations where there are multiple types of documents that need to be evaluated.

    Set up a document classification use case in the following steps.
    1. Create a document classification use case.

      Define the name and properties for the use case.

    2. Create a document class.

      Define the classes or categories that the AI will learn to detect and apply to documents.

    3. Create a document task for document classification and upload sample documents for each class.
    4. Train a use case.

      Initiate a training job to provide user inputs from completed document tasks to the AI for continuous improvement.