Skip to main content

​
How Knowledge Documents Power Analysis

Knowledge Documents are the fuel for Fabius’s contextual understanding. They actively inform and guide how Fabius analyzes sales interactions and automates related tasks.

​
1. Providing Context via AI Selection

When Fabius analyzes an event (like a call or email), it doesn’t just look at the interaction in isolation. It first determines which of your Knowledge Documents are relevant to that specific conversation.
  • How it Works:
    1. Fabius looks at a Generated Summary (made by an AI when we ingest the document) of each available, completed Knowledge Document version.
    2. It compares these summaries to the content of the current interaction (e.g., call transcript, email thread).
    3. An AI model identifies the documents whose summaries most closely relate to the topics discussed in the interaction.
  • Outcome: Fabius dynamically selects a small subset of your most relevant Knowledge Documents to use as context for the current analysis task.
The Power of Summaries: This selection process highlights why accurate, AI-generated Internal Summaries are vital. Good summaries allow Fabius to quickly and correctly identify the most relevant knowledge for any given situation. You don’t need to manually tag documents for relevance; Fabius figures it out based on content.

​
2. Guiding Analysis with Enumeration Fields

Knowledge Documents integrate deeply with Enumeration Fields (the building blocks of your analysis configurations) through Knowledge Document Prompt Instructions.

Linking Documents to Analysis Fields

When configuring an Enumeration Field, you can link relevant Knowledge Documents to it (using the “Knowledge Documents” section on the field’s settings page). This tells Fabius that a particular document contains information pertinent to that specific analysis point.

Setting the 'Usage Mode'

Each link has a “Usage Mode” that dictates how the document influences the field’s analysis:
- Reference during analysis: The processed content of the linked document is included directly in the prompt given to the AI when analyzing this field for an interaction. This mode is used when the document contains procedures, definitions, or examples the AI should consult directly. Condition: The document must also have been selected as relevant to the overall interaction by the AI Selector (Step 1).
- Update Prompt: The document’s content is not included in the live analysis prompt. Instead, this mode enables automatic maintenance. If the linked document is updated (a new version is created with changes), Fabius can use the summary of changes to suggest updates to the Enumeration Field’s Instructions, Directions, or Definitions, keeping your analysis logic aligned with your latest knowledge.
- Both: The document is used for both direct context during analysis and for triggering potential prompt updates when the document changes.

​
Example Scenario:

  1. Setup: You link your “Standard Discovery Questions” Knowledge Document to the “Discovery Score” Enumeration Field with Usage Mode set to Reference during analysis.
  2. Interaction: A sales call occurs where the prospect discusses budget early on.
  3. AI Selection: Fabius’s AI Selector identifies the “Standard Discovery Questions” document as relevant because it contains sections on budget qualification.
  4. Field Analysis: When Fabius calculates the “Discovery Score” for that call, it includes the content of the “Standard Discovery Questions” document in the prompt sent to the scoring AI.
  5. Result: The AI can now evaluate how well the rep followed the specific discovery questions outlined in your document when scoring their performance, leading to a more accurate and contextually relevant score.

​
3. Driving Automation (e.g., Email Drafts)

Knowledge Documents are essential for ensuring automated outputs align with your specific processes and messaging.
  • Email Drafting: When Fabius generates an email draft (because your Gmail is integrated):
    • It selects relevant Knowledge Documents (e.g., an FAQ for the product that the customer is asking about, or a playbook generated from historical emails that is similar to the current situation).
    • The content of these documents provides context to the AI, helping it structure the email correctly, include appropriate next steps, mention relevant resources (like case studies referenced in a linked document), and maintain your company’s tone.
  • Task Generation: If an analysis identifies that a specific procedure from a Knowledge Document needs to be followed (e.g., “Escalate technical issue to Tier 2 support”), Fabius can generate a task based on the steps outlined in that document.
By curating a comprehensive set of Knowledge Documents, you directly influence the quality, relevance, and accuracy of Fabius’s analysis and automation, making it a powerful tool tailored specifically to your way of doing business.