Workflow Strategy: AI-Assisted "Forms Over Data" Development

💡 The Core Strategy: "The Golden Master"

This workflow relies on Reference Implementations. By hand-coding the first set of screens (The "Golden Master"), you create perfect training data for the AI. You then treat the AI as a smart scaffolder to replicate that pattern across other entities, drastically reducing repetitive work while maintaining architectural control.

🏗️ Phase 1: Foundation (Hand Build Only)

Goal: Establish architectural constraints, security, and UI standards. AI should inhabit the architecture, not invent it.

• 1. App Shell & Navigation: Hand build to ensure the UX frame is solid.
• 1a. Security & Login: Hand build. Never outsource auth logic to AI without heavy scrutiny.
• 2. Database Access: Hand build the setup/ORM configuration to ensure patterns (Repository, Unit of Work) are correct.
• 3. Entities: Hand build throughout the SDLC. Your domain model is the business truth; AI often hallucinates relationships.
• 4. Notification System: Hand build the pattern (Interfaces for Email/SMS/Teams) to ensure reliability.

⚡ Phase 2: Replication (AI Assisted)

Goal: High velocity on repetitive tasks using the "Golden Master" from Phase 1 as a prompt reference.

5. Browse Screens

• Workflow: Build the first one (e.g., CustomerBrowse) by hand. Use AI to generate ProductBrowse, OrderBrowse, etc.
• AI Instruction: "Create a browse screen for [Entity] using CustomerBrowse as the exact visual and functional template."

6. Query & Export Screens