SECI-GRAI Knowledge Creation Framework

This skill provides the theoretical foundation for understanding knowledge creation cycles, particularly in human-AI collaboration contexts. It integrates Nonaka and Takeuchi's SECI model with the modern GRAI (Generative, Receptive AI) extension.

Core Concept: Knowledge Types

All knowledge exists on a spectrum between two forms:

The creation of new organizational knowledge occurs through continuous conversion between these types.

The SECI Model: Four Conversion Modes

Knowledge creation follows a spiral through four modes:

1. Socialization (Tacit → Tacit)

What it is: Sharing tacit knowledge through shared experiences, observation, imitation, and practice.

Indicators present phase is Socialization:

• Learning by watching someone work
• Pair programming or shadowing
• Informal knowledge transfer ("let me show you how")
• Building shared mental models through collaboration
• Apprenticeship-style learning

Key activities:

• Joint problem-solving sessions
• Collaborative exploration of a domain
• Sharing war stories and experiences
• Building rapport and shared understanding

AI-Human pattern (GRAI):

• Human→AI: Iterative prompting with rich contextual information
• AI→Human: Explaining topics, demonstrating approaches, walking through reasoning

2. Externalization (Tacit → Explicit)

What it is: Articulating tacit knowledge into explicit concepts—the most critical and difficult conversion.

Indicators present phase is Externalization:

• Documenting how something works
• Writing specifications from understanding
• Creating diagrams, models, or frameworks
• Explaining "why" decisions were made
• Converting intuition into guidelines

Key activities:

• Writing documentation from experience
• Creating product specifications
• Defining processes and workflows
• Building conceptual models
• Articulating design rationale

AI-Human pattern (GRAI):

• Human→AI: Adding materials via memory/context to refine understanding
• AI→Human: Converting unstructured knowledge into structured formats

3. Combination (Explicit → Explicit)

What it is: Combining, categorizing, and systematizing explicit knowledge into new forms.

Indicators present phase is Combination:

• Synthesizing multiple documents
• Building knowledge bases or wikis
• Creating summaries from various sources
• Restructuring existing documentation
• Cross-referencing and linking concepts

Key activities:

• Merging multiple specifications
• Creating comprehensive guides from fragments
• Building taxonomies and categorizations
• Generating reports and dashboards
• Systematizing best practices

AI-Human pattern (GRAI):

• Human→AI: Using AI creatively to combine unlikely patterns
• AI→Human: Generating summaries, meeting protocols, synthesis documents

4. Internalization (Explicit → Tacit)

What it is: Embodying explicit knowledge through learning-by-doing until it becomes tacit.

Indicators present phase is Internalization:

• Learning from documentation
• Practicing new skills
• Applying guidelines in real situations
• Building muscle memory and intuition
• "Making it your own"

Key activities:

• Hands-on practice with documented procedures
• Simulations and exercises
• Applying patterns to new contexts
• Building intuition through repetition
• Developing personal heuristics

AI-Human pattern (GRAI):

• Human→AI: AI observing patterns to suggest timely support
• AI→Human: Supporting human understanding, creating practice exercises

The Knowledge Spiral

Knowledge creation is not linear but spiral—each cycle builds on the previous:

        Socialization ──────► Externalization
              ▲                      │
              │                      ▼
              │    KNOWLEDGE         │
              │      SPIRAL          │
              │                      │
        Internalization ◄────── Combination
              │                      ▲
              └──────────────────────┘
                   (next cycle)

Spiral dynamics:

• Each cycle expands the knowledge base
• Individual knowledge becomes team knowledge becomes organizational knowledge
• The spiral moves through different social levels (individual → group → organization)

GRAI: The AI Extension

The GRAI framework (Generative, Receptive AI) extends SECI for human-AI collaboration by recognizing AI as an active participant in knowledge creation.

Eight Interaction Fields

GRAI doubles the SECI phases by adding direction (human↔machine):

Human-Centered Design

GRAI maintains human agency through two configurations:

• Human-in-the-loop: Human makes decisions, AI augments capability
• Machine-in-the-loop: AI handles routine work, human provides oversight

The framework preserves human decision-making authority while leveraging AI for knowledge work amplification.

Phase Identification Quick Reference

To identify the current phase, ask:

Applying SECI-GRAI

For Documentation Work

For Product Development

Phase Transition Triggers

Moving between phases often requires deliberate action:

Common Pitfalls

Skipping Externalization: Trying to combine knowledge that hasn't been articulated yet results in shallow synthesis.

Premature Combination: Combining sources before deeply understanding them produces surface-level results.

Neglecting Socialization: Pure documentation without shared experience lacks the tacit context that makes knowledge actionable.

Incomplete Internalization: Reading without practice leaves knowledge as information, not capability.

Additional Resources

Reference Files

For detailed theory and advanced applications, consult:

• references/seci-deep-dive.md - Complete Nonaka & Takeuchi theory with academic foundations
• references/grai-framework.md - Full GRAI framework details and interaction patterns
• references/phase-transitions.md - Techniques for facilitating movement between phases

Example Files

Working examples in examples/:

• phase-identification-examples.md - Real-world scenarios with phase analysis

Integration with Other Skills

This skill provides the theoretical foundation. Related skills in knowledge-manager:

• ba-contexts - Enabling contexts for each SECI phase
• knowledge-assets - Types of knowledge artifacts to create
• extension-interface - Patterns for tool-specific implementations

Tool-specific plugins (e.g., km-notion, km-obsidian) extend these foundations with platform-specific patterns.