Paradigms of Generative Artificial Intelligence in Automating Corporate Code Writing

Аннотация

This paper examines the paradigm shifts in SUBMITED 19 July 2025 ACCEPTED 29 July 2025 leveraging generative artificial intelligence for PUBLISHED 12 August 2025 automated code generation at the enterprise level. It is VOLUME Vol.07 Issue 08 2025 thus a critical review of prevailing prescriptions for CITATION integrating LLM agents into the software development lifecycles of modern enterprises, assessing their impact . on team productivity and the new risks they introduce COPYRIGHT © 2025 Original content from this work may be used under the terms to confidentiality and licensing matters. The study would of the creative commons attributes 4.0 License. therefore be most befitting at this stage, as fast-forward steps are being made towards organizational adoption of generative AI, from mere IDE autocompletion features to more than a co-programmer but an autonomous agent capable even of popping pull requests sans humans in the loop, demanding new forms of legibility both organizationally and technically. The novelty of this research lies in its integration of material from scholarly works, industry reports, and case studies, along with lab pilot runs of Copilot and actual DevSecOps implementations, to triangulate the current state and future promise of this technology on a practical business level. Key findings include: a reduction of development cycle time by 50–60% without compromising code quality thanks to the integration of AI agents into IDEs and CI/CD pipelines; a shift of developers’ roles toward architects and reviewers as routine tasks are delegated to digital co‑programmers; and a necessity for phased implementation that accounts for private code protection and compliance with licensing norms. Significant barriers identified include model hallucination management, ensuring the traceability of changes, and adapting organizational culture and regulations to new roles such as prompt designers and AI-agent curators. The article will be of use to IT department heads, software architects, The American Journal of Engineering and Technology 92 https://www.theamericanjournals.com/index.php/tajet DevSecOps specialists, and researchers in the field of humans focus on requirements specification, artificial intelligence. architecture validation, and critical-change review, while the mundane tasks are handed over to digital Keywords: generative artificial intelligence, coprogrammers, opening up new reserves of speed, programming automation, corporate SDLC, AI agents, quality, and flexibility for organizations. DevSecOps Materials and Methodology