Step 1: Create a Project

From your Avylo dashboard, select New Project. Give your project a clear name and initial scope.

Step 2: Provide Meaningful Project Knowledge

Paste or write a rich description of what you are building. Great inputs include:

• The core problem your product solves and for whom.
• Key user flows (e.g., “Users upload large video files, which are transcribed and indexed for semantic search”).
• Expected scale and traffic characteristics (e.g., “Targeting 10,000 daily active users with bursty morning workloads”).
• Known constraints (e.g., “Must run on AWS, require SOC 2 compliance within 6 months”).

Step 3: Avylo Analyzes the Project

Once submitted, the AI Architect decomposes your project knowledge:

1. Formulates the Blueprint.
2. Discovers business Domains and Capabilities.
3. Derives functional and non-functional Requirements.
4. Drafts baseline Architecture Decisions.

Step 4: Resolve Material Uncertainty

Avylo only interrupts you for material business or architecture uncertainty. If confidence is high on standard patterns (e.g., relational storage for structured billing data), Avylo automatically forms documented assumptions.

If a critical fork exists (e.g., synchronous vs asynchronous processing for long-running AI pipelines), Avylo will present focused questions to clarify intent.

Step 5: Bootstrap Discovery Completion

When overall discovery confidence reaches the target threshold (typically ≥ 82%), discovery automatically completes. Avylo generates your Architecture V1 and transitions you directly into the persistent Architecture Workspace.