Integration of Generative Al in Enterprise Software: Enhancing Workflow Automation And Decision-making

Abstract

This thesis offers a structured literature review on the integration of generative artificial intelligence (AI) into enterprise software systems, following academic recommendations to adopt a research-focused approach. It synthesizes insights from over 100 peer-reviewed articles, white papers, and standards published between 2018 and early 2025. The study explores four core themes: (1) integration architectures and design patterns, including microservices, serverless, and hybrid cloud environments; (2) enterprise applications such as chatbots, automated reporting, document summarization, scenario forecasting, and anomaly detection; (3) governance and ethical considerations, focusing on bias mitigation, explainability, and data privacy; and (4) emerging trends like foundation models, low-code development, multimodal AI, and environmentally sustainable AI metrics. The research finds that modular, API-centric architectures are widely adopted for embedding AI into legacy systems. While enterprises benefit from increased automation and operational efficiency, challenges persist—such as vendor lock-in, skills shortages, data quality issues, and weak governance structures. The thesis concludes by identifying research gaps in sustainable AI, human–AI collaboration, and model lifecycle management, offering practical recommendations for responsible AI integration in enterprise contexts.