Document ID: AOD-006 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
Required By:
- Governance Workflow
Purpose
Discovery is the first operational phase of every AOD project.
Its objective is not to design software.
Its objective is to build a complete understanding of the business before engineering begins.
No architecture, implementation, or governance documentation should be created until Discovery has been completed.
Discovery Objectives
The AI should leave Discovery with a clear understanding of:
- Why the organization exists
- The business problem
- The desired outcome
- The target users
- The customers
- Success criteria
- Business processes
- Constraints
- Risks
- Existing systems
- Terminology
- High-level product vision
If significant uncertainty remains, Discovery should continue.
Discovery Topics
Discovery should naturally progress through the following areas.
The AI should adapt the conversation to the human rather than rigidly following a script.
Organization
Understand:
Mission
Vision
Industry
Business model
Competitive advantage
Current stage of growth
Problem
Understand:
What problem exists?
Who experiences it?
Why is it important?
How is it solved today?
Why does the current solution fall short?
Desired Outcome
Understand:
What does success look like?
How will success be measured?
What changes after success is achieved?
Users
Identify:
Primary users
Secondary users
Administrators
External participants
Stakeholders
Business Processes
Understand:
Current workflows
Pain points
Manual activities
Automation opportunities
Decision points
Constraints
Identify:
Budget
Timeline
Technology preferences
Regulatory requirements
Security expectations
Platform requirements
Risks
Identify:
Technical risks
Business risks
Operational risks
Adoption risks
Scalability risks
Vocabulary
Capture:
Important business terms
Industry terminology
Role definitions
Product terminology
Acronyms
Words with special meaning
AI Behavior During Discovery
The AI should:
Ask one logical question at a time.
Keep conversations conversational.
Summarize frequently.
Verify assumptions.
Challenge ambiguity.
Recognize conflicting information.
Avoid discussing implementation details unless required for clarification.
Focus on understanding rather than solving.
Discovery Completion Criteria
Discovery is complete when the AI can confidently explain:
The business.
The problem.
The desired outcome.
The users.
The major business processes.
The primary constraints.
The success criteria.
Any remaining assumptions.
The AI should present this summary to the human for confirmation.
Discovery is not complete until the human agrees that the summary accurately represents the business.
Discovery Deliverable
The output of Discovery is not documentation.
The output is a shared understanding between the human and the AI.
Once confirmed, the AI transitions to the Governance phase.
No governance documents should be generated until Discovery has been successfully completed.
Transition to Governance
After Discovery has been approved, the AI should announce:
"Discovery Complete."
The AI should summarize the agreed understanding and recommend beginning Governance.
Governance transforms the shared understanding into structured engineering documentation that will guide the remainder of the project.