1. Applications Administration Guide
  2. Use Case 3: Multimodal AI

Use Case 3: Multimodal AI

In this scenario, Generative AI integrates and processes multiple types of data such as text, and images—to generate more comprehensive insights and responses. This bridges the gap between vision and language, interpreting the context, and delivering results in a contextually relevant response.

Use Case Scenario 1: Closing or Transfer Summary

After completing all the work associated with a Service Request and issue is resolved, Agent can generate a Closing Summary. Transfer Summary for another Agent is generated in case Agent needs to transfer Service Request to another Agent.

  • Create new activity and select Type=Closing Summary / Type=Transfer Summary.
  • Drawer window displays the Closing / Transfer Summary.
  • User will have an opportunity to review and edit response before saving Closing / Transfer Summary.
Note:
  1. After a new Closing / Transfer Summary activity is selected, Siebel will trigger call to the configured Large Language Model (LLM) to generate a Closing / Transfer Summary based on all inbound, outbound communications and Service Request Attachments (images)
  2. Currently, only the last uploaded Attachment(image) in the Service Request Attachment Applet is sent to the Large Language Model (LLM)

  3. The file formats supported are JPEG, JPG, PNG

Use Case Scenario 2: Creating a Service Request

While entering the necessary fields associated with a Service Request, Agent can use AI model to to automatically detect predefined values, extract text from documents uploaded and auto-populate in the respective field
  • In the Service Requests list, create a new record and save
  • In the Attachments applet, click Add File / Add URL
  • Click Extract button
  • Pre defined values are auto-populated in the respective fields
  • User will have an opportunity to review and edit before saving
Note:
  1. After Extract button is clicked, Siebel will trigger call to the configured Large Language Model (LLM) to extract text, and other key data from the image
  2. Currently, you can pass only a single image per Large Language Model (LLM) call for extraction
  3. The file formats supported are JPEG, JPG, PNG

Configuration and Authentication

To configure Generative AI capabalities

  • To enable and configure Generative AI capabilities, set the following System Preference:

    Name Vaue

    EnableOCIAIGenAI

    N (to be changed to 'Y' to activate feature)
    Note: By default it is disabled.

  • Create seed data, as required (using the database setup wizard or through manual configuration)

To authenticate Generative AI

  • Download the Authentication file from OCI which has the Credential details like userId, fingerprint, and tenancy.
  • Rename the file Config to ociconfig.properties and place it in Siebel's internal Tomcat's webapps folder.
  • Ensure that the ociconfig.properties file includes the location of the private key (.pem) file.
  • Import the necessary certificates as required to both internal and external Tomcats and add it to the Siebel Keystore and Truststore.