International Journal of Science, Engineering and TechnologyAuthors: Dr. Alexander Hayes, Dr. Emily Carter, Daniel Foster, Dr. Sophia Reynolds, Michael Bennett, Jeji Krishnan
Abstract: The exponential growth of log data in distributed enterprise systems has made traditional monitoring and manual root-cause analysis increasingly inefficient and error-prone. This paper presents a generative AI–driven framework for automated log summarization and root-cause triage, enabling faster and more accurate diagnosis of system failures. The proposed approach leverages large language models to transform unstructured and high-volume log streams into concise, context-aware summaries, while simultaneously identifying anomalous patterns and correlating events across distributed components. By integrating evidence mapping techniques with AI-driven diagnostics, the framework establishes a unified view of system behavior, significantly reducing the cognitive load on support engineers. Additionally, the study explores the use of retrieval-augmented generation and feedback loops to continuously improve model accuracy and adaptability in dynamic environments. Empirical evaluation across enterprise-scale platforms demonstrates notable improvements in incident triage time, reduction in mean time to resolution, and enhanced operational efficiency. The findings highlight the potential of generative AI to transform enterprise observability, shifting from reactive troubleshooting to intelligent, automated, and scalable root-cause analysis in complex distributed systems.
DOI: http://doi.org/10.5281/zenodo.20265420
