Insider Brief
- Abeja and Yamazaki Mazak have begun a joint development project focused on physical AI for factory planning, machining, production scheduling and robotic operations.
- The companies will use Abeja Platform to develop AI agents for machining-process design and production scheduling, along with a training environment for robots and material-handling systems.
- The systems are intended to learn from factory operating data and worker knowledge over time, with the longer-term goal of integrating them into Yamazaki Mazak machine tools and automation systems.
Abeja and machine-tool maker Yamazaki Mazak have begun a joint development project to develop physical AI for more automated factory planning, machining and robotic operations.
According to Abeja, the Japanese companies will use Abeja’s AI platform to develop software agents that can help design machining processes and production schedules, along with a system for training robots and other automated equipment. The longer-term goal is to incorporate the technology into Yamazaki Mazak machine tools and automation systems used by its manufacturing customers.
The work targets manufacturing processes that still depend heavily on experienced workers, particularly process planning, production scheduling and equipment setup. The companies said those areas are becoming more difficult to manage as manufacturers contend with worker shortages and the loss of expertise as skilled employees leave the workforce.
AI for Machining and Production
The project uses Abeja Platform, the company’s system for developing and operating AI in enterprise settings. Abeja describes its approach as “human in the loop,” with workers remaining involved as AI systems learn and make operational recommendations.
The joint development work will focus on three areas.
Machining processes — The companies plan to develop AI agents that can propose how a part should be manufactured, including machining methods, processing sequences and cutting conditions. ABEJA said the system is intended to capture some of the decision-making knowledge traditionally held by experienced technicians and make it available across manufacturing operations.
Production scheduling — A second set of AI agents will generate and adjust production schedules using information including production plans, equipment availability, component inventories and operating knowledge from factory workers. The companies said the system could help manufacturers respond more rapidly to disruptions such as component shortages or urgent orders.
Robot training — Abeja and Yamazaki Mazak will also build an environment for training robots and material-handling systems. The platform will collect operating data from factories, train and evaluate AI models and test their performance on physical equipment.
Abeja said the systems will continue learning from operating data and worker knowledge after deployment, allowing them to adapt to individual factories and production environments.
