Explanation: Why CMMN for Knowledge Work

This document explains the rationale behind using CMMN for knowledge-intensive work and SEA™’s design decisions.

1. Why CMMN?

1.1 The Knowledge Work Problem

Traditional PM (Waterfall/Agile) assumes prescriptive workflows:

• Fixed phases/sprints
• Predefined tasks
• Linear/iterative progression

Knowledge work is discretionary:

• Tasks emerge based on findings
• Progression depends on events
• Non-linear, adaptive workflows

Example:

1
2
3
4
Research Project:
- Can't plan all experiments upfront
- Hypothesis may change mid-project
- Additional analysis emerges from results

1.2 CMMN’s Discretionary Model

CMMN (Case Management Model and Notation) provides:

2. Artifact Pipeline Rationale

2.1 The Reification Progression

Knowledge work produces artifacts at different abstraction levels:

1
Cognitive (notes) → Intellectual (specs) → Product (deliverables) → Capital (assets)

Why Four Stages?

1. Cognitive: Reduce mental load (ephemeral, personal)
2. Intellectual: Synthesize knowledge (project-scoped, team-shared)
3. Product: Deliver value (market/domain-facing)
4. Capital: Organizational asset (enterprise-wide, reusable)

2.2 Provenance Gating

Problem: Artifacts appear without lineage (“teleportation”)

Solution: TransitionTokens enforce progression:

1
2
3
Can't create Product without Intellectual
Can't create Capital without Product
Every transition logged to IFL

Benefit: Immutable lineage enables:

• Audit trails
• Quality gates
• Capital valuation
• Knowledge reuse

3. PM-Agent Design Decisions

3.1 Why AI Orchestration?

Human PM Challenges:

• Overwhelmed coordinating specialists
• Manual task routing
• Context switching overhead
• Difficulty tracking dependencies

PM-Agent Solution:

• Automated routine orchestration
• Capability-based assignment
• Continuous sentry evaluation
• Knowledge Base integration

3.2 Hybrid Teams

Why Not All-Human or All-AI?

Design: Humans focus on judgment, AI handles execution, PM-Agent coordinates.

4. Trade-Offs and Alternatives

4.1 CMMN vs Waterfall

Trade-off: Accept less predictability for better adaptability.

4.2 CMMN vs Agile

Trade-off: Less ceremony, more discretion.

4.3 Artifact Pipeline vs Ad-Hoc

Trade-off: Accept pipeline overhead for capital formation.

5. Future Evolution

5.1 Planned Enhancements

• Predictive sentries: ML models suggest task activation
• Cross-case learning: PM-Agent learns from past cases
• Automated capitalization: AI identifies capital candidates
• Portfolio optimization: Balance cases for maximum capital formation

5.2 Open Questions

• How to handle highly uncertain work (pure research)?
• How to balance discretion with accountability?
• How to value Intellectual Capital objectively?

6. Conclusion

CMMN provides a proven, standards-based model for discretionary workflows. SEA™’s adaptation adds:

• Artifact pipeline for knowledge capitalization
• PM-Agent for AI orchestration
• Provenance gating for immutable lineage

By treating knowledge work as event-driven and discretionary, organizations achieve flexibility, artifact reuse, and measurable capital formation.