How can AI agents be used reliably in a wide range of manufacturing settings to increase both productivity and sustainability?
Generative AI agents are already proving themselves in areas like coding and writing. But moving from digital applications to reliable real-world factory operations is a different challenge altogether. An agent might analyse production data, investigate a recurring fault or suggest an intervention. But can it work reliably within operating constraints? Can it recognise when data is incomplete or outdated? And does it know when to stop and ask for human input?
These are some of the questions that motivated our work on the AI-enabled Manufacturing and Maintenance (AIMM) Framework.
I'm pleased to share our new whitepaper, developed through our work at VITO, together with Ivo Adan and Nitish Singh from Eindhoven University of Technology, and shaped by valuable contributions from industrial partners in Brainport and Flanders.
Manufacturers are working with existing infrastructure and large amounts of digital information, while facing skills shortages and pressure to use energy and materials efficiently. These conditions create both an opportunity and a challenge for AI agents working alongside humans to improve productivity and sustainability.
Our central idea is that reliable agentic AI requires more than better models. It requires a systems-engineering approach that brings together infrastructure, meaningful data, scalable and secure services, and clear responsibilities and knowledge-sharing between humans and agents.
That foundation matters for both productivity and sustainability. It helps manufacturers improve equipment performance, reduce energy peaks, and support more circular use of resources.
We bring these elements together in a four-layer framework, illustrated through an investigation of Overall Equipment Effectiveness (OEE). The same approach, along with investments in the underlying infrastructure and services, can support further applications, including production scheduling, equipment maintenance, and quality improvement.
Many thanks to everyone who contributed to the discussions and development of this work, and to Brainport Industries and European Supply Chain Forum - ESCF for publishing it.
More about the AI-enabled Manufacturing and Maintenance (AIMM) Framework
If you would like to know more about this whitepaper and how AI agents can be used reliably in a wide range of manufacturing settings to increase both productivity and sustainability, feel free to contact Shane.