# How I built an AI tool to stop guessing software architecture from bad requirements

> Source: <https://dev.to/ingenierofard/how-i-built-an-ai-tool-to-stop-guessing-software-architecture-from-bad-requirements-2h57>
> Published: 2026-09-14 16:08:17+00:00

Hey everyone! 👋

We've all been there. A client or product manager hands you a 2-paragraph document or a chaotic PDF and says: *"We need a simple Uber for X, please give us a technical estimation and the architecture by tomorrow."*

You spend the next few hours scratching your head, guessing the database entities, mapping potential modules, and praying you didn't miss a massive technical blind spot that will ruin the project two months from now.

I got so tired of this "improvisation phase" that I decided to build a solution for it: **Noesor** (noesor.com).

Noesor is an AI-powered platform designed around the **Design-First** philosophy. Its core mission is simple: transform raw, ambiguous text requirements or PDF documents into a clear, structured technical foundation **before** you write a single line of code.

Instead of jumping blindly into coding, it acts as an intelligent bridge to help you make well-founded architectural decisions.

When you input a requirement, the contextual engine processes the text and automatically extracts:

`User`, `Transaction`, `Booking`) and maps how they relate to each other.
As a developer, I hate signing up for tools just to see if they actually work. That's why I designed Noesor with a **strict no-friction onboarding**:

I built this to solve my own pain points, but now I want to make it awesome for the global dev community.

Go to noesor.com, throw your most chaotic, poorly written requirement at it, and tell me:

Drop your thoughts, feedback, or roast in the comments below! 👇