Impact of Artificial Intelligence on Modern Software Development Lifecycle: Methodological Advancements, Developer Productivity, and Quality Engineering

The modern Software Development Lifecycle (SDLC) is undergoing a profound paradigm shift driven by the rapid maturation of generative artificial intelligence, Large Language Models (LLMs), and automated program synthesis engines. While historical transitions—from linear Waterfall models to Agile sprints and continuous DevOps pipelines—focused primarily on procedural orchestration and organizational agility, the integration of Artificial Intelligence introduces cognitive automation directly into the engineering fabric. This research paper presents an exhaustive empirical and methodological investigation into the impact of AI across the end-to-end SDLC, evaluating its tangible ramifications on developer productivity, cognitive workflow dynamics, code quality engineering, and enterprise security. Employing a rigorous triangulated mixed-methods research design, this study combines: (i) quantitative repository mining across 1,200,000 commits spanning 45 enterprise software projects; (ii) a controlled randomized quasi-experiment comprising N = 120 professional software engineers measuring task velocity and cognitive workload via the NASA-TLX inventory; and (iii) longitudinal static analysis tracking post-release defect density and Common Weakness Enumerations (CWE). Our findings demonstrate that AI assistance compresses coding lead times by 55.8% and accelerates developer flow states, yielding substantial improvements across DORA deployment frequency and change velocity metrics. However, ungoverned AI code generation introduces acute quality paradoxes, including a 19.4% surge in code churn, subtle logic regression risks, and elevated vulnerabilities in insecure deserialization and unverified package dependencies. To resolve this tension, we propose a multi-tier Enterprise AI-Augmented SDLC (A-SDLC) Framework coupling Abstract Syntax Tree (AST) policy guardrails, local context retrieval-augmented generation (RAG), automated mutation test synthesis, and continuous AIOps observability. Finally, we analyze organizational governance imperatives, intellectual property licensing risks, and emerging horizons in autonomous multi-agent coding swarms, establishing an empirical blueprint for resilient, high-assurance software engineering.