AI is driving the smart factory forward, but Italian businesses are struggling to keep up
Companies are implementing individual projects without addressing the issue as part of a long-term strategy. Artificial intelligence offers opportunities for development, but the data is not always used effectively
Artificial intelligence is knocking on the doors of Italian factories. But not all of them are ready to open them. This is the picture that emerges from the latest report by the Internet of Things Observatory at the Politecnico di Milano: a mixed picture of an industry that recognises the potential of new technologies but is struggling to translate this into strategy.
The figures tell the story of two industrial Italies. Among large companies, 71 per cent have launched at least one Industrial IoT project, comprising sensors, connected machinery and monitoring platforms. Among medium-sized enterprises, this figure drops to 59 per cent. But the most significant gap concerns the long-term vision: only 29 per cent of large companies and just 6 per cent of medium-sized ones have established multi-year programmes involving several departments.
“There remains a significant gap in terms of the industrial strategies underpinning the projects,” note the researchers at the Politecnico. The majority of initiatives – 41 per cent among large firms and 69 per cent among medium-sized firms – remain confined to ‘one-off’ measures: specific projects focusing on individual production lines or machines, without an overarching strategy.
The paradox is that Italian factories are producing ever-increasing amounts of data, but often do not know what to do with it. No medium-sized company reports carrying out advanced processing of the information generated by its IoT systems: not even a quarter of large companies do so, whilst the rest merely use the data ‘in its raw form’ or, worse still, leave it unused.
When data is actually used, it is predominantly for diagnostic purposes: to monitor performance and determine when to carry out maintenance. Predictive analytics, which aims to anticipate faults, optimise consumption and forecast demand, is used by only a third of active companies. Adaptive use – which enables systems to automatically adjust processes in real time – remains marginal: just 4 per cent.
