AI agent combines OT expertise and generative AI to support factory maintenance.
Daikin Industries, Ltd. and Hitachi, Ltd. have begun trial operations of an AI agent designed to support equipment failure diagnostics in factories. The trial started in April 2025 at Daikin’s Sakai Plant-Rinkai Factory in Sakai City, Osaka Prefecture, where commercial air conditioning systems are produced. The initiative aims to commercialize the “AI Agent for Equipment Failure Diagnostics” by integrating long-standing operational technology (OT) expertise with advanced generative AI.
The AI agent is developed to assist maintenance technicians in identifying the causes of equipment failures—such as in pumps or valves—and suggesting corrective actions using a tablet interface. It leverages Daikin’s accumulated factory equipment drawings and converts them into machine-readable knowledge graphs. These are then combined with maintenance records and Hitachi’s proprietary failure analysis process based on STAMP methodology, forming a dataset for AI learning.
In a preliminary demonstration, the AI agent identified failure causes and recommended solutions in under 10 seconds with an accuracy exceeding 90%. This performance is equivalent to or better than that of standard maintenance technicians at Daikin.
Background
In addition to a decline in the working-age population in the manufacturing sector, the number of skilled technicians has also decreased. Meanwhile, globalization and the expansion of production bases have increased the demand for maintenance technicians, exacerbating labor shortages.
Daikin operates over 90 production sites in 28 countries and continues to expand globally. However, some overseas factories face challenges in hiring and training skilled maintenance personnel, particularly as manufacturing equipment becomes more advanced. These challenges have led to longer maintenance lead times and inconsistent maintenance quality. Although data such as equipment logs and maintenance records have been collected, they have not yet been fully utilized to enhance operational efficiency.
With the workforce shrinking and social challenges becoming increasingly complex, there is an urgent need to reduce the burden on frontline workers and improve productivity in manufacturing environments. In response, Hitachi established the Generative AI Center in May 2023 to support the secure and effective deployment of generative AI. The company aims to drive productivity gains through its Lumada solutions, leveraging synergies between IT, OT, and deep domain knowledge gained through collaborations with manufacturing clients.
Daikin and Hitachi have previously collaborated on next-generation production models, including digitalizing brazing work, fluorochemical reaction processes, and optimizing production and sales planning in the chemicals sector. The companies now seek to deepen their partnership to realize a continuously evolving, zero-downtime factory.
“By combining OT knowledge from the field with advanced IT, we’re enabling diagnostics that improve both speed and precision,” said the companies in a joint statement.