Processes
Explore lifecycle processes, operational workflows, and governance reviews that guide how teams build and operate across the Velocity Blueprint.
Lifecycle processes (xDLCs)
Operational processes
Governance processes
Agent Development Life Cycle (ADLC)
End-to-end lifecycle for building and deploying AI agents.
Lifecycle Agentic core
Traditional
1
2
Design
Design agent architecture and prompts
Responsible: AI/ML architect (AIA)
Architecture reviewed
Components:
3
4
5
6
AI-Native
Spec & Planning Planning Agent
Human role: Specifies intent declaratively
Outputs:
Architecture decisionsComponent selectionAmbiguity flags
Eval Design (TDD) Eval Agents
Human role: Defines success criteria
Outputs:
Eval datasetsAccuracy benchmarksRegression tests
Coding Coding Agents
Human role: Reviews direction
Outputs:
Agent codePromptsMCP toolsGuardrails
Local Validation Eval Agents
Human role: Iteratively tests in workspace
Outputs:
Local eval runsPrompt tuningBehavior validation
Release Train Review & Deploy Agents
Human role: Monitors pipeline
Outputs:
Security checksCanary deployArchitectural validation
Maintain & Refactor Refactoring Agents
Human role: Reviews suggestions periodically
Outputs:
Tech debt cleanupPrompt optimizationDependency updates