Document ID: AOD-010 Version: 1.0.0 Status: Approved
Depends On:
- AOD-001 Manifest
- AOD-002 Core Principles
- AOD-003 AI Responsibilities
- AOD-004 Human Responsibilities
- AOD-005 Initialization Prompt
- AOD-006 Discovery Workflow
- AOD-007 Governance Workflow
- AOD-008 Architecture Workflow
- AOD-009 Planning Workflow
Required By:
- Validation Workflow
Purpose
The Implementation phase transforms the approved implementation plan into working software.
Implementation should execute the approved engineering plan rather than redefine it.
The objective is disciplined execution, not continued product discovery.
Objectives
Implementation should:
Produce working software.
Follow the approved architecture.
Remain aligned with governance documentation.
Protect maintainability.
Minimize technical debt.
Support future scalability.
Continuously validate progress.
Implementation Principles
Implementation should:
Follow the approved roadmap.
Build only approved scope.
Keep changes small.
Validate frequently.
Document significant implementation decisions.
Protect existing functionality.
Prefer clarity over cleverness.
AI Responsibilities
During Implementation the AI should:
Follow approved governance.
Protect the established architecture.
Implement one logical unit of work at a time.
Explain major implementation decisions.
Recommend improvements when appropriate.
Identify risks before implementation.
Refuse implementation requests that violate approved architecture without first discussing the consequences.
Update governance documents when implementation reveals previously unknown information.
Maintain implementation consistency throughout the project.
Human Responsibilities
During Implementation the human should:
Review completed work.
Provide timely feedback.
Approve milestone completion.
Clarify changing business needs.
Evaluate tradeoffs.
Avoid introducing unnecessary scope changes.
Change Management
New ideas will naturally emerge during development.
When a proposed change significantly impacts architecture, scope, timeline, or business objectives, the AI should pause implementation and recommend returning to the appropriate AOD phase.
Examples include:
Major feature additions
Architecture changes
Business model changes
Security model changes
Platform changes
Significant workflow changes
Small implementation improvements may proceed without restarting the methodology.
Major changes should never bypass governance.
Continuous Validation
Implementation should be continuously evaluated against:
Approved governance documents
Approved architecture
Implementation roadmap
Business objectives
Success criteria
The AI should immediately identify deviations and recommend corrective actions.
Completion Criteria
Implementation is complete when:
All approved implementation objectives have been achieved.
The approved MVP or release scope has been completed.
Known defects have been addressed or documented.
Governance documentation reflects the implemented solution.
The product is ready for formal validation.
Stage Gate
Before transitioning to Validation the AI should present:
Implementation Summary
Completed Features
Deferred Features
Known Issues
Outstanding Risks
Validation Readiness Assessment
Implementation should not be considered complete until the human has approved the completed work.