Configuring Content Assistant to use enhanced annotations

You can configure Content Assistant to use enhanced text extraction annotations from Content Platform Engine 26.0.0 or later for improved document processing.

Before you begin

Make sure that Content Platform Engine 26.0.0 or later is deployed in your environment with enhanced extraction configured. For more information about enhanced extraction, see Enhanced extraction overview External link opens a new window or tab.

About this task

Enhanced extraction in Content Platform Engine 26.0.0 provides advanced text extraction capabilities that process complex document types such as PDF forms, images, and structured data. Enhanced extraction can produce document content in two formats:

  • Markdown format - Document content rendered in Markdown format, which is the preferred format for vector indexing and generative AI queries when available.
  • Plain text format - Document content in plain text, which is combined with extracted key-value pairs and table data when available.

You can configure Content Assistant to specify which format to use for different operations: vector indexing, document summarization, and search queries.

Procedure

  1. In the Administration Console for Content Platform Engine (ACCE), access the Gen AI Configuration object for the object store where you want to configure enhanced annotations.
    1. In the left navigation pane, expand your object store.
    2. Expand Administrative > Content Assistant.
    3. Click Gen AI Configuration to open the object.
  2. Configure the annotation format for vector indexing.
    1. Click the Vector Indexing tab.
    2. Select or clear the Prefer markdown extract checkbox to control the annotation format that is used when populating the vector index.

      When the checkbox is selected, Content Assistant uses the markdown format. When the checkbox is cleared, Content Assistant uses the plain text format that is combined with key-value pairs and table data if available.

      Note: The markdown format is recommended when markdown annotations are available, as it typically provides better results for AI operations.
  3. Configure the annotation format for document summaries.
    1. Click the Summary Generation tab.
    2. Select or clear the Prefer markdown extract checkbox to control the annotation format that is used when generating document summaries.

      When the checkbox is selected, Content Assistant uses the markdown format. When the checkbox is cleared, Content Assistant uses the plain text format that is combined with key-value pairs and table data if available.

  4. Configure the annotation format for search and user queries.
    1. Click the Search tab.
    2. Select or clear the Prefer markdown extract checkbox to control the annotation format that is used when processing user queries and generating AI responses.

      When the checkbox is selected, Content Assistant uses the markdown format. When the checkbox is cleared, Content Assistant uses the plain text format that is combined with key-value pairs and table data if available.

  5. Click Save to apply your changes.

Results

The Gen AI Configuration object is updated with your preferred annotation format settings.