We’ve been experimenting with automating the "first month" of a web project—the tedious phase of setting up repos, configuring Tailwind, building basic components, and arguing over spacing. By generating the design assets and the code simultaneously from a shared requirement set, we managed to condense a typical 8-12 week setup foundation into roughly 3-10 hours.
It’s not perfect, and the designs won't win awards for originality, but for B2B dashboards, it feels like we might be done writing boilerplate manually.
The core technical problem we hit wasn't generating code (AI is decent at that now); it was the "translation drift." Usually, a designer hands off a static Figma file, and a developer translates it into dynamic code. They drift apart almost immediately. Most AI tools just generate snippets or single screens, which leaves you with a "Frankenstein" UI—inconsistent tokens, varied padding, and no shared architecture.
We realised the issue was trying to sync two different sources of truth. So we tried a "System-First" approach.
Instead of Design → Code, we built a pipeline (we call it TheSSS AI) that works like this: Requirements → Context-Aware Interpretation → Governance.
The AI generates the rules first (tokens, accessibility standards, spacing scales) in JSON. Then, it spawns the artefacts in parallel from that single truth:
- Design Tokens (JSON)
- UI Mockups (Visual assets)
- Component Library (React/Vue + Storybook)
- Documentation (Markdown)
Because the React button component and the Figma mockup are generated from the same parent data at the same time, there is no translation error. They just match.
The results are interesting. We can generate a full "Foundation-Ready" MVP—Auth flows, Dashboard layouts, Settings pages, basic CRUD views—in about a working day. The system handles the mundane stuff nobody on our team likes doing: accessibility annotations, responsive variants, and writing the documentation files.
But there are significant trade-offs, and I want to be honest about them here:
The "SaaS Look": The designs are clean and accessible, but they look like "Standard Bootstrap/SaaS." If you need a unique, award-winning brand identity, this approach feels too rigid. It effectively automates the "boring" layer of design, not the creative layer.
The Logic Gap: It generates the UI and state management boilerplate perfectly, but it doesn't understand deep business logic. It can build the "Edit User" modal, but it won't know the complex validation rules for your specific legacy database unless you explicitly prompt for every edge case (which takes longer than just coding it).
The "Eject" Problem: This is the big open question for us. Regenerating the entire foundation is easy, but what happens three months in? Once a human developer takes over and modifies the code, you can't easily "re-run" the AI without overwriting manual changes. We're still figuring out the reconciliation strategy here.
I'm curious if others are experimenting with this "simultaneous generation" approach? Or are we moving toward a world where the initial codebase is disposable and we just regenerate it when requirements change?
Does anyone else feel like the "setup" phase of software development is becoming obsolete, or does automating it just encourage code bloat?
Top comments (0)