Overall Equipment Efficiency (OEE) is a comprehensive metric used in manufacturing and industrial environments to measure the effectiveness of equipment and processes. It evaluates how well a manufacturing operation is utilized compared to its full potential, factoring in three critical elements: availability (the percentage of scheduled time that the equipment is ready to operate), performance (the speed at which the equipment runs compared to its designed capacity), and quality (the proportion of good products produced versus the total output). By combining these factors, OEE provides a single, actionable score that highlights opportunities for improving productivity, minimizing downtime, and reducing waste. Organizations rely on OEE to identify bottlenecks, track progress, and guide continuous improvement strategies within their plants. Organizations face numerous data-related challenges when analyzing OEE, such as fragmented and siloed data across different departments and systems, inconsistent and unstructured data formats from sources like free-text maintenance notes or incident reports, and difficulties in collecting real-time data due to outdated equipment or manual processes that introduce latency. The reliability of OEE insights is further undermined by data quality issues, such as sensor errors and manual entry mistakes and by the ongoing struggle to integrate legacy systems not built for modern data analytics. Moreover, the absence of standardized definitions for key OEE components (such as downtime or quality losses) can make cross-plant benchmarking unreliable, while growing concerns over data security and privacy add further complexity as organizations grapple with regulatory requirements and access controls. Overcoming these obstacles requires coordinated efforts in data integration, quality assurance, automation, and the adoption of advanced solutions like AI-powered agents and knowledge graphs to ensure that OEE metrics truly drive operational improvement. Identifying the root causes of poor OEE currently requires the analysis of various plant data, such as predictive maintenance records, work orders, timesheets, root cause analyses (RCAs), and event logs. However, manually going through these large and often unstructured datasets is a complicated, time-consuming, and potentially error-prone task, which delays the identification and resolution of OEE issues. Additionally, the interaction of multiple factors can obscure the underlying cause of performance problems. Our research focuses on how to leverage agentic AI and knowledge graphs to enhance OEE. Specifically, the method optimizes OEE by developing AI agents based on Large Language Model (LLMs) that leverage a comprehensive knowledge graph. This knowledge graph represents a digital twin of the plant, including equipment specifications, performance history, failure records, and work order details. Recent research has demonstrated the synergistic potential of LLMs when combined with knowledge graphs. By integrating LLMs with explicit, structured knowledge representations, such as those found in industrial or scientific settings, researchers have achieved notable improvements in both the factual accuracy and interpretability of AI-driven analysis. Hybrid systems now leverage LLMs’ natural language understanding alongside the relational depth of knowledge graphs, enabling sophisticated querying, precise information retrieval, and advanced reasoning for complex, real-world domains.