Skip to main content

​
Kodexa Knowledge

Kodexa Knowledge is the third foundational pillar of the Kodexa platform, complementing Studio and Workflow to provide a comprehensive and adaptive document processing ecosystem. It formalizes and centralizes operational intelligence, enabling intelligent automation and dynamic adaptation to diverse business requirements.

​
The Challenge: Disconnected Management

In previous versions of Kodexa, managing content understanding was a disjointed experience. Critical configurations—such as defining document types, setting up extraction rules, managing taxonomies, and orchestrating workflows—were handled through separate, often disparate interfaces. This fragmentation led to:
  • Operational inefficiencies
  • Increased complexity for administrators
  • Difficulty adapting to evolving business needs
  • Challenges in implementing consistent governance
  • Complex integration of new capabilities

​
Completing the Platform

Kodexa Knowledge brings everything together, forming the third pillar:

​
Studio: Unified Design & Configuration

Studio serves as the central hub where everything comes together, offering a unified experience for building, testing, and managing all core document processing components.

​
Workflow: Operational Command Center

Workflow orchestrates execution of document processing pipelines, managing tasks and data flow at scale with enhanced capabilities for creating projects, tasks, and searching documents.

​
Knowledge: The Adaptive Intelligence Layer

Knowledge empowers solutions with context-aware logic and dynamic adjustments, critical for evolving business rules and intelligent automation.

​
Three Foundational Concepts

Kodexa Knowledge is built on three core concepts that work together to provide intelligent, adaptive document processing:

​
1. Knowledge Types

Developer-defined templates that establish the structure and expected data types for specific knowledge entities. Knowledge Types are foundational templates, typically provided by Kodexa or developers, that define the schema for information you want to extract and organize. They come with out-of-the-box defaults and can be customized using transformers. Example Use Cases:
  • Specifying criteria for ignoring certain data elements based on predefined classifications
  • Configuring custom text prompts for data elements
  • Defining parameters for document processing behavior
By defining Knowledge Types, Kodexa gains the context needed to intelligently route, enrich, and process documents.

​
2. Knowledge Features

Business-aligned attributes that are observable, extractable, and serve as fundamental data points for decision-making. Knowledge Features are the business-aligned attributes captured about your documents, projects, or tasks during processing. They range from simple properties to complex identifiers: Simple Examples:
  • Document Type (string)
  • Approval Status (boolean)
  • Effective Date (date)
  • Page Count (number)
Complex Examples:
  • Transaction Identifier (multiple fields)
  • Vendor Information (composite data)
These features are configured in Kodexa Studio, as their implementation often requires code linking to define extraction logic and behavior.

​
3. Knowledge Sets

Conditional applications that enable dynamic processing based on specific feature values. Knowledge Sets are collections of rules and associated actions that activate based on specific feature values, allowing for dynamic adjustments in processing. Each Knowledge Set has:
  • Name: Descriptive identifier (e.g., “Specific Process Automation Rule”)
  • Status: Active, Pending Review, On Hold, or Archived
  • Conditions: Feature-based triggers
  • Actions: Processing adjustments to apply
Example: 
When Document Type = 'Invoice' AND Value > '1000'
  → Apply additional approval workflow
The “Pending Review” status is designed for future AI suggestions, where the system can detect patterns and recommend new rules.

​
Knowledge Application Levels

Kodexa Knowledge offers flexible scope management for applying operational intelligence:

​
Organization Level

Knowledge applied universally across an entire organization, ensuring consistent operational intelligence for all operations and departments.

​
Project Level

Knowledge scoped to specific projects, where different initiatives may have unique processing requirements.

​
Hybrid Approach

Knowledge Sets can be applied to “one or more projects” or “all projects,” combining broad applicability with targeted adjustments. This comprehensive approach provides unparalleled flexibility, enabling Kodexa to adapt to a wide spectrum of business models and processing requirements.

​
Knowledge Lives in Documents

Knowledge is applied during document processing and becomes part of the document metadata:
  • Applied During Processing: Knowledge rules are evaluated and applied as documents move through pipelines
  • Visible in Review: Applied knowledge is visible when reviewing documents
  • Searchable: Documents can be searched and filtered by applied knowledge
  • Traceable: You can see exactly which knowledge rules affected each document
  • Integrated with Workflow: Knowledge integrates seamlessly with Workflow’s document search capabilities
The Knowledge interface provides cards similar to Studio and Workflow for consistency, with Knowledge Sets, Documents, and Reporting as the main sections.

​
Operational Flow

Kodexa Knowledge follows a systematic operational flow:

​
1. Contextual Understanding

Extracts and interprets critical information, transforming raw data into actionable insights that drive operational decisions for every document.

​
2. Dynamic Decisioning

Automatically applies the most relevant rules and logic based on document context, ensuring precise, consistent, and compliant operational outcomes.

​
3. Adaptive Workflow Orchestration

Intelligently adjusts processing paths and actions in real-time, optimizing workflows for maximum efficiency, accuracy, and reduced operational costs.

​
4. Intelligent Automation

Delivers highly accurate and context-aware processing, minimizing manual intervention and significantly improving overall operational excellence.

​
Benefits

​
Scalable Operational Processing

Handles increasing diversity and volume of document types and business rules without escalating system complexity or custom coding.

​
Simplified Management & Maintenance

Reduces the administrative burden and technical overhead associated with managing evolving business logic and processing rules across operations.

​
Dynamic & Adaptive Flexibility

Enables automatic adjustment of processing behavior based on real-time document context, ensuring agility and reducing manual intervention.

​
Orchestrating Dynamic Processing

Knowledge Sets eliminate redundant processing pipelines and static configurations by:
  • Modifying Requirements: Adjusting data extraction field requirements based on conditions
  • Customizing Instructions: Tailoring processing instructions for specific data elements
  • Applying Validation: Enabling or disabling validation rules based on document characteristics
  • Making Document-Level Decisions: Enhancing agility and control at the individual document level
This feature-driven architecture streamlines operations across organization-wide or project-specific applications.

​
Future Vision

​
AI-Driven Knowledge Discovery (Coming 2026)

The “Pending Review” status enables future AI-driven knowledge discovery: Workflow:
  1. AI Detects Patterns: Identifies recurring operational patterns and validation exceptions
  2. Generates Suggestions: Creates new knowledge set suggestions based on detected patterns
  3. Pending Review: Sets suggested knowledge sets to “Pending Review” status
  4. Expert Approval: Human experts review and approve suggested knowledge
  5. Activate & Automate: Approved knowledge becomes active, enhancing automated processing
Example Use Case: “This type of exception always gets ignored for vendor ID X” → AI creates a knowledge set suggestion → Operations team reviews and activates → Exception is automatically handled going forward This transforms reactive exception handling into proactive knowledge management.

​
Knowledge-Driven Autonomous Agents

Building on Kodexa Knowledge, the future roadmap includes Agentic Dataflows—autonomous agents that leverage accumulated knowledge to make context-aware decisions about:
  • Processing steps
  • Validation requirements
  • Data enrichment
  • Workflow routing
These agents will be guided by sophisticated logic and rules captured within Knowledge Types, Features, and Sets, and continually refined by human experts, enabling truly intelligent and operations-centric processing.

​
Getting Started

Knowledge is configured and managed through three main interfaces:
  1. Studio: Define Knowledge Types and Features, configure extraction logic
  2. Knowledge Interface: Create and manage Knowledge Sets, review applied knowledge
  3. Workflow: See knowledge in action during document processing and review
For implementation guidance and examples, refer to the Concepts section on Options and the Cookie Cutters for practical templates.

​
Summary

Kodexa Knowledge completes the platform alongside Studio and Workflow, providing the adaptive intelligence layer that enables:
  • Context-aware document processing
  • Dynamic rule application
  • Intelligent automation
  • Scalable operational management
  • Foundation for future AI-driven capabilities
This integration ensures a seamless journey from solution design to execution and intelligent adaptation, transforming how Kodexa handles complex and varied document processing requirements.