A Service Agent at Universal Containers (UC) is designed to help customers resolve issues by searching against knowledge articles. Knowledge articles have PDF attachments that add critical details…

Question

Options

• AIncrease article chunk size and token limits for Knowledge indexing so larger contexts capture
• BEnable 'Include Related Attachments' for Knowledge-kav and map the ContentDocumentLink
• CUse Data Cloud's `Include Attachments' option and select the ContentDocumentVersion

• A
  17% (8)
• B
  8% (4)
• C
  75% (36)

Explanation

When Knowledge articles have PDF or file attachments that contain critical information, the Data Cloud search index must be explicitly configured to ingest that file content - it does not happen automatically. In Data Cloud's setup for Knowledge (Knowledge__kav) object indexing, there is a specific 'Include Attachments' toggle that, when enabled, allows you to select the ContentDocumentVersion object. ContentDocumentVersion is the Salesforce object that holds the actual binary/text content of an attached file. Selecting it tells Data Cloud to extract and index the textual content from those files (including PDFs), making them discoverable via vector search. Option A (increasing chunk size/token limits) only affects how already-ingested text is split - it does not pull in new content from attachments. Option B references 'Include Related Attachments' and ContentDocumentLink, which is not the correct Data Cloud configuration path; ContentDocumentLink is a junction object that links documents to records but is not what Data Cloud uses to extract file content.

No community discussion yet for this question.