The factory of the future between AI, data and organisational culture

The digital transformation of industry is undergoing a decisive transition: after years in which innovation focused mainly on management systems and planning and progress control tools, the focus is now shifting directly to the heart of production. Connected machines, continuous data flows and Artificial Intelligence technologies put at the service of their processing for optimisation, promise a new season for manufacturing, to make it even more resilient and competitive. But the path is not a linear one and companies find themselves moving in a working environment characterised by increasing pressures.

Margins are shrinking, global competition is growing, regulations are becoming more stringent and the difficulty in finding qualified personnel is slowing down generational change within production departments. Meanwhile, customers demand ever smaller batches, greater customisation, increasing quality and service, as well as processes that reduce waste and environmental impact. It is not only the market that is changing, but the very way in which a factory must be designed and governed. For many companies, understanding how to exploit the potential of digital and AI has become a key strategic element.

In recent years there has been - as mentioned - a significant spread of digital tools to support planning and scheduling and process control (MES), areas where technologies and interoperability are relatively mature. However, it is much less common to see real use of machine-generated data. And yet, it is precisely there - in the signals collected by sensors and PLCs on board machines - that an enormous amount of information is hidden, capable of improving the quality, efficiency and continuity of processes if appropriately analysed and processed to derive useful indications on how to govern the process or manage and calibrate a machine. The unfamiliarity with this data combined with the technologies and tools to extract and process it, together with regulations that favour the connection of plants without requiring the effective exploitation of the information collected, still holds many companies back.

In this scenario, Artificial Intelligence represents a concrete opportunity. Machine Learning, Deep Learning and generative models are no longer abstract concepts, but operational tools that find application in many areas of production. From waste prediction to process optimisation, from predictive maintenance to real-time support for operators, in avatar mode, AI can make the factory more stable and robust. Some systems are even able to automatically adjust machine parameters, reducing variability and deriving efficiency without human intervention. A possibility that until a few years ago belonged to the realm of experimentation and is now within the reach of even small and medium-sized companies.

Surprising to many is the fact that these projects can be implemented in a relatively short time and with limited investment, one or two orders of magnitude less than that required for a new machine. This makes the adoption of AI particularly attractive even in traditional production contexts and in small to medium-sized companies.