Skip to main content

Core Concepts

Before managing Knowledge Documents, it’s helpful to understand the key components:

1. Knowledge Document

This is the main container entity representing a single piece of knowledge (e.g., “Q3 Sales Playbook”, “Competitor Battlecard - Acme Inc.”). Key attributes you’ll interact with:
  • Name: A user-friendly identifier you provide.
  • Description: A brief summary of the document’s purpose.
  • Type: Categorizes the document based on its function. Examples: “Marketing Materials”, “Product Documentation”, “Sales Training Materials”. This helps Fabius understand how to potentially use the information.
  • Source: Indicates how the document originated in Fabius. Examples: “Manual Upload”, “Google Docs”, “Generated from Calls”.
  • Current Version: Points to the latest content and metadata associated with this document.

2. Knowledge Document Version

Documents change. Fabius tracks these changes through versions. A new version is automatically created every time the content of a Knowledge Document is updated (e.g., uploading a new file, editing text in the Fabius editor, reprocessing from a URL). Editing only metadata like the Name or Description does not create a new version. Each version stores important historical information:
  • Version Number: An identifier (1, 2, 3…).
  • Content: Both the original (raw) content and the cleaned-up (processed) text used by Fabius are stored securely.
  • Source Metadata: Details specific to how this version was created. Examples:
    • Google Docs: The specific Document ID and Revision ID from Google at the time of import.
    • Webpage: The URL it was fetched from and when.
    • Generated from Calls: Which representative’s calls were used.
    • Generated from Emails: Which user’s emails and the specific module name extracted.
    • Videos: Details like the Video ID, Title, and Author.
  • Processing Status: Shows whether Fabius has finished processing the content for this version. For example, if you upload a long video, it may take a few minutes to process. Text is generally processed immediately.
When a new version is created, its content needs to be processed (cleaned) and summarized by Fabius. This happens automatically in the background. You’ll see the status update in the UI:
StatusMeaningWhat it means for you
PendingQueued for processing.Content not yet ready for use.
ProcessingActively being processed (cleaning text, generating summary).Content not yet ready for use. Usually takes a few minutes.
CompletedSuccessfully processed and summarized.Ready! The content and summary can now be used by Fabius analysis.
FailedAn error occurred (e.g., couldn’t read file, AI summary failed).Check the error message. Content likely unusable. May need re-upload.
UnknownAn unexpected state.Might indicate an issue. Contact support if persistent.
Only versions marked as Completed are used by Fabius for providing context during analysis or automation. If you see the content of the document in the UI, then it’s in the Completed state.
  • History: Easily track how important documents like playbooks or product specs have changed.
  • Auditability: See when updates were made.
  • Context Integrity: Ensures analyses are based on the document version that was current at that time.
  • Reference: View content from older versions if needed.

3. Document Sources

This identifies where the document came from. Different sources have slightly different creation steps and capture different metadata:
SourceDescriptionHow You Add ItKey Metadata Captured (Examples)
Manual UploadYou upload a file (PDF, TXT, MD, DOCX, audio, video…).”Upload Files” tab.File type (MIME).
Customer WebpagesFabius scrapes content from a public website URL.”Import from the Web” tab.URL, Time Pulled.
Google DocsFabius imports content from a Google Doc (requires GSuite integration).”Import from the Web” tab (paste GDoc URL).Google Doc ID, Revision ID, URL.
Generated from CallsFabius analyzes calls from a selected top performer to create a best practices doc.”Create from Calls” tab.Rep Name, Specific Calls Used.
Generated from EmailsFabius analyzes historical emails to extract specific procedures or playbooks.Triggered via a dedicated action (e.g., “Analyze Emails”).User Analyzed, Module Name.
Videos (e.g., YouTube)Fabius fetches the transcript from a video URL.”Import from the Web” tab (paste Video URL).Video Title, Author, URL.
UnknownFallback or placeholder.Shouldn’t typically occur in normal use.None.
Understanding these core concepts helps you manage your knowledge base effectively. Next: Creating & Managing Documents