# Unified Architecture Metamodel of Information Systems Developed by Generative AI

arXiv: https://arxiv.org/abs/2604.00171 · HTML: https://mlsanity.com/paper/2604.00171/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 28 references that are in the map.
- First posted on arXiv 31 Mar 2026.

## Builds on (most cited first, 9 of 9)

- [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://mlsanity.com/paper/2310.03714/) (2023)
- [Software Architecture Meets LLMs: A Systematic Literature Review](https://mlsanity.com/paper/2505.16697/) (2025)
- [How Scientists Use Large Language Models to Program](https://mlsanity.com/paper/2502.17348/) (2025)
- [Capturing Software Architecture Knowledge for Pattern-Driven Design](https://mlsanity.com/paper/2005.08393/) (2020)
- [Sketches and Diagrams in Practice](https://mlsanity.com/paper/1706.09172/) (2017)
- [Generating Software Architecture Description from Source Code using Reverse Engineering and Large Language Model](https://mlsanity.com/paper/2511.05165/) (2025)
- [Towards Living Software Architecture Diagrams](https://mlsanity.com/paper/2407.17990/) (2024)
- [Training Software Engineers for Qualitative Evaluation of Software Architecture](https://mlsanity.com/paper/2105.09595/) (2021)
- [An Explanation of Software Architecture Explanations](https://mlsanity.com/paper/2503.08628/) (2025)

## Built on by (most cited first, 4 of 4)

- [Specification-Based Code-Text-Code Reengineering for LLM-Mediated Software Evolution](https://mlsanity.com/paper/2605.25232/) (2026)
- [Token Optimization Strategies for LLM-Based Oracle-to-PostgreSQL Migration](https://mlsanity.com/paper/2605.28557/) (2026)
- [Measuring What a Specification Determines: A Formal Semantic-Block Model and an Execution-Judged Benchmark](https://mlsanity.com/paper/2608.19475/) (2026)
- [Specification Portability Across LLM Development Agents: Cross-Agent Compatibility in Specification-Driven Software Migration](https://mlsanity.com/paper/2608.21208/) (2026)

## Abstract (arXiv)

The rapid development of AI and LLMs has driven new methods of SDLC, in which a large portion of code, technical, and business documentation is generated automatically. However, since there is no single architectural framework that can provide consistent, repeatable transformations across different representation layers of information systems, such systems remain fragmented in their system representation. This study explores the problem of creating a unified architecture for LLM-oriented applications based on selected architectural frameworks by SMEs. A framework structure is proposed that covers some key types of architectural diagrams and supports a closed cycle of transformations, such as: "Code to Documentation to Code". The key architectural diagrams are split equally between main architectural layers: high-layer (business and domain understanding), middle-layer (system architecture), and low-layer (developer-layer architecture). Each architectural layer still contains some abstraction layers, which make it more flexible and better fit the requirements of design principles and architectural patterns. The conducted experiments demonstrated the stable quality of generated documentation and code when using a structured architectural context in the form of architectural diagrams. The results confirm that the proposed unified architecture metamodel can serve as an effective interface between humans and models, improving the accuracy, stability, and repeatability of LLM generation. However, the selected set of architectural diagrams should be optimised to avoid redundancy between some diagrams, and some diagrams should be updated to represent extra contextual orchestration. This work demonstrates measurable improvements for a new generation of intelligent tools that automate the SDLC and enable a comprehensive architecture compatible with AI-driven development.