Skip to content

Running Inception

The Inception phase turns a project description into structured requirements, user stories, and tasks. It is the first phase of the AI-DLC methodology lifecycle.

Starting Inception

Navigate to your sprint and write a project description:

  • For a greenfield project (new, from scratch): describe the system, its goals, constraints, and scope
  • For a brownfield project (existing codebase): describe the feature, issue, or change you want to make

Choose Launch Agent to start the Inception Agent.

Before the agent runs, the system performs a readiness check on your description.

Readiness check

The readiness check has two parts:

Structural checks

These are fast, deterministic checks:

  • Minimum content length
  • Required information present
  • Enough detail for the agent to work with

AI analysis

If structural checks pass, the LLM reviews the description for:

  • Clarity and completeness
  • Ambiguous requirements
  • Missing technical details
  • Feasibility

You see the results in real time as the AI streams its analysis. If the check fails, you get specific feedback about what to improve.

Bypassing readiness

If you believe the description is ready despite failing the readiness check, you can choose to proceed anyway.

Generated artifacts

After Inception completes, you see:

  • Requirements with titles, descriptions, and acceptance criteria
  • User stories with descriptions and story points
  • Tasks mapped to concrete implementation work

All artifacts are editable. You can refine them before moving to Construction.

Questions

During Inception, the agent may ask clarifying questions. These appear in the UI as structured prompts with suggested answers. Answer them to help the agent produce better artifacts.

Moving to Construction

Once you are satisfied with the generated artifacts, select a git branch and base branch, then launch the Construction Agent. The sprint transitions to the Construction phase.