How I Built This Blog with AI-DLC: A New Way to Develop Software with AI

What if you could have an experienced software architect guiding you through every step of building an application? That’s exactly what AI-DLC offers - and I used it to build this very blog you’re reading.

What is AI-DLC?

AI-DLC (AI Development Lifecycle) is a structured framework for AI-assisted software development. Think of it as a methodology that transforms how AI models like Claude collaborate with developers on real projects.

Instead of ad-hoc prompting and hoping for the best, AI-DLC provides:

• Structured phases with clear checkpoints
• Adaptive execution based on project complexity
• Complete audit trails of all decisions
• Human-in-the-loop approval at every critical step

The Three Phases

AI-DLC organizes development into three distinct phases:

┌─────────────────────────────────────┐
│         INCEPTION PHASE             │
│    Planning & Architecture          │
│                                     │
│  • Workspace Detection              │
│  • Requirements Analysis            │
│  • User Stories                     │
│  • Workflow Planning                │
│  • Application Design               │
└─────────────────────────────────────┘
                  ↓
┌─────────────────────────────────────┐
│       CONSTRUCTION PHASE            │
│   Design, Build & Test              │
│                                     │
│  • Functional Design                │
│  • NFR Requirements                 │
│  • Infrastructure Design            │
│  • Code Generation                  │
│  • Build and Test                   │
└─────────────────────────────────────┘
                  ↓
┌─────────────────────────────────────┐
│        OPERATIONS PHASE             │
│    Deploy & Monitor                 │
│                                     │
│  • Deployment Planning              │
│  • Monitoring Setup                 │
│  • Maintenance Workflows            │
└─────────────────────────────────────┘

INCEPTION: What to Build and Why

The Inception phase answers the fundamental questions: What are we building? Why? For whom?

For this blog, the AI-DLC workflow:

1. Detected this was a greenfield project (starting from scratch)
2. Gathered requirements through structured questions about purpose, features, and constraints
3. Created user stories defining key personas (Reader, Subscriber, Admin)
4. Planned the workflow deciding which stages to execute
5. Designed the application architecture with components and services

Output: Clear requirements, user personas, 7 epics, and a complete application design - all before writing a single line of code.

CONSTRUCTION: How to Build It

The Construction phase transforms designs into working software:

1. Code Generation Plan - A detailed checklist of what to create
2. Code Generation - Actual implementation of all components
3. Build and Test - Verification that everything works

For this blog, Construction produced:

• 6 Astro pages (Home, Blog, About, Services, Contact, RSS)
• React components for interactive features
• AWS Amplify backend configuration
• Cognito authentication setup
• DynamoDB data models
• Complete styling with Tailwind CSS

OPERATIONS: Deploy and Run

The Operations phase (currently a placeholder for future expansion) will handle:

• Deployment to AWS Amplify
• Monitoring and observability
• Incident response procedures

How I Built This Blog

Here’s the actual journey through AI-DLC:

Step 1: Initial Request

I started with a simple prompt:

“Using AI-DLC, build a blog for my domain agiusalexandre.com. I’m an AWS Solutions Architect, expert on AWS, passionate about AI & Security.”

Step 2: Requirements Gathering

AI-DLC asked me structured questions:

• Purpose? Brand building + tutorials + portfolio
• Content type? Mix of long-form and short posts
• Content management? Markdown files in Git
• Hosting? AWS Amplify Gen2
• Features? Full-featured (posts, comments, newsletter, search, RSS)
• Languages? English + French (i18n)

Each answer shaped the requirements document.

Step 3: User Stories

AI-DLC generated personas and epics:

Personas:

• Alexandre (Content Creator)
• Professional Reader
• Casual Visitor
• Newsletter Subscriber

Epics:

1. Core Blog Infrastructure
2. Content Discovery & Navigation
3. Visitor Engagement
4. Professional Presence
5. Content Subscription
6. Content Management
7. Analytics & Optimization

Step 4: Application Design

AI-DLC designed:

• 5 Components: BaseLayout, PostCard, ContactForm, NewsletterForm, SearchBar
• 4 Services: ContentService, AuthService, SubscriptionService, AnalyticsService
• 3 Data Models: Comment, Subscription, ContactMessage

Step 5: Code Generation

With the approved design, AI-DLC generated all code files:

• Astro pages and layouts
• React components
• Amplify backend configuration
• Styling and configuration files

Step 6: Build and Test

Finally, AI-DLC verified the build:

npm run build
# Result: 6 pages built successfully

The Power of Iteration

Here’s where AI-DLC really shines: iteration after the first build.

Greenfield vs Brownfield

AI-DLC distinguishes between:

• Greenfield: New projects starting from scratch
• Brownfield: Existing codebases with established patterns

Once your initial build is complete, subsequent changes are treated as brownfield - meaning AI-DLC understands the existing architecture and adapts.

Adaptive Execution

Not every change needs every stage. AI-DLC intelligently assesses what’s needed:

Change Type	Stages Executed
New feature	Full workflow
Bug fix	Minimal (Code Gen + Test)
Content addition	Streamlined (Code Gen only)
Architecture change	Comprehensive

This Blog Post is an Example

This very blog post was created using AI-DLC in iteration mode:

1. Workspace Detection: Recognized existing codebase
2. Requirements Analysis: Minimal - clear request
3. Workflow Planning: Streamlined approach
4. Code Generation: Create markdown file
5. Build and Test: Verify build

The framework didn’t make me go through full requirements gathering and user stories for a simple blog post. It adapted to the task.

Continuous Improvement Loop

┌─────────────────────────────────────────────────────┐
│                                                     │
│   INCEPTION ──→ CONSTRUCTION ──→ OPERATIONS         │
│       ↑                              │              │
│       │                              │              │
│       └──────── Feedback ←───────────┘              │
│                                                     │
│   Each iteration:                                   │
│   • Leverages existing architecture                 │
│   • Skips unnecessary stages                        │
│   • Maintains audit trail                           │
│   • Builds on previous decisions                    │
│                                                     │
└─────────────────────────────────────────────────────┘

Why This Matters

Traditional AI-assisted coding often feels chaotic:

• Prompts that work once but not twice
• Context lost between sessions
• No record of why decisions were made
• Inconsistent code quality

AI-DLC solves these problems:

Traditional AI Coding	AI-DLC
Ad-hoc prompts	Structured phases
Lost context	Session continuity
No audit trail	Complete documentation
Inconsistent	Repeatable process
Developer-driven	AI-guided with human approval

Getting Started with AI-DLC

Want to try AI-DLC? Here’s how:

1. Setup

Add the AI-DLC framework files to your project:

.kiro/
├── steering/aws-aidlc-rules/
│   └── core-workflow.md
└── aws-aidlc-rule-details/
    ├── common/
    ├── inception/
    └── construction/

2. Create CLAUDE.md

Add a CLAUDE.md file to integrate with Claude Code:

# AI-DLC Integration

For software development requests, follow:
.kiro/steering/aws-aidlc-rules/core-workflow.md

3. Start Building

Trigger the workflow with:

"Follow the AI-DLC workflow to [your task]"

The Tech Stack

For reference, here’s what AI-DLC helped me build:

Layer	Technology
Frontend	Astro 5 + React 19
Styling	Tailwind CSS
Backend	AWS Amplify Gen2
Auth	AWS Cognito
Database	DynamoDB
Hosting	AWS Amplify + CloudFront

Conclusion

AI-DLC represents a paradigm shift in how we collaborate with AI on software projects. Instead of treating AI as a code autocomplete tool, it becomes a structured development partner that:

• Guides you through proven methodologies
• Adapts to your project’s needs
• Maintains accountability and traceability
• Enables confident iteration

This blog is living proof: built from scratch using AI-DLC, and continuously improved using the same framework.

The future of software development isn’t AI replacing developers - it’s AI augmenting developers with structured, repeatable, and adaptable workflows.

---

Interested in learning more about AI-DLC or my AWS Solutions Architecture work? Get in touch or explore more posts.