There is no longer any doubt that artificial intelligence is redesigning production paradigms. Yet, in the Italian industrial fabric, its adoption still struggles to transform itself into a real lever for widespread growth.
This is confirmed by a photograph taken by the Digital Innovation Observatory in SMEs of the Polytechnic University of Milan, which highlights an emblematic fact to say the least: just 4% of small and medium-sized businesses have started structured training courses on AI.
To learn more with Panorama the cultural and operational barriers that hold back SMEs, the role of the experiential approach and the concrete challenges of technology transfer in companies, was Miriam SnackDirector of MADE Competence Center.
According to the Polytechnic Observatory, only 4% of SMEs have started structured training on AI. What are the main barriers that hold back small and medium-sized Italian businesses? Is it a question of economic resources, lack of awareness or fear of technological complexity?
I believe that the first obstacle is to contextualize the investment: before asking how much a training course costs, an entrepreneur must be clear about what will change in his company, in terms of work organization and productivity. AI may still seem far from the daily needs of business, something that does not concern the immediate future. It is a perception that needs to be changed, rather than a question of tools.
On this ground, the training group of the MADE Competence Center, led by Marta Rispoli, is actively engaged. It is a work of dissemination and orientation of great importance, in my opinion: without it, even support such as tenders and funding dedicated to the sector risk remaining unused, because there is a lack of awareness of their usefulness. Even the fear of technological complexity often arises from this lack of awareness: understanding the potential of AI helps to identify where to start.
Then, there are practical constraints, starting with the time to dedicate to training. I wouldn’t dismiss them as a lack of focus on growth: in a small business, freeing up even a few people from day-to-day tasks is a challenging choice. This is why I feel that our responsibility, as MADE, is to make the benefits of professional development recognisable: to make them perceived not only as a cost, but rather as an investment that is measured on two levels, the human one – people feel involved and the company retains its best resources – and the economic one, which the entrepreneur needs to see translated into concrete effects on the business.
Starting from a real difficulty, making people try, leaving room for questions, even the simplest ones: this is the approach of the centers of expertise. Financing certainly helps. But the willingness to learn arises when a person sees a possibility for growth that concerns them too, not just a change that everyone talks about. This applies to top management, as well as direct operators.
How does SkillAI actually work? How does AI interview employees and ensure that skills mapping is accurate and not perceived as a “punitive” assessment test?
SkillAI starts from a question, rather than a pre-packaged answer. Born from the collaboration between MADE and the Polytechnic of Milan, it offers an interview that adapts to what the worker reports. If he says «I use AI to prepare a document», the system elaborates: «For what activity? How did I set up the request? What did I check before using the result?” It is a model divided into seven dimensions, from the individual’s understanding of AI to operational use, up to the verification of outputs and awareness of risks. The answers are then read and interpreted, highlighting fragilities and ways of improvement.
Of course, a single interview, no matter how well planned, does not capture everything a person can do, and caution is essential. The company organization must also be considered: which tools the operator can actually use, with which rules, with how much space to experiment. To avoid it feeling like an exam, you need to be clear about the purpose and use of the information collected. I really recognize myself in the idea of a map: something that helps guide learning, leaving the possibilities for growth open. A useful return should make you think “now I understand where to start”, not “this is all I know how to do”.
Mapping skills is the first step, but what happens next? How does the information collected by SkillAI translate into personalized and truly effective training plans for business growth?
A map is useful if it helps you choose a route. After SkillAI, therefore, the question that needs to be asked is: what is worth learning to do better? The information collected must be related to the company’s objectives, internal processes and available tools. It is from this comparison that a personalized path is born, to be built with the company. Training, in this sense, is not a standard intervention, applicable to every situation: it is a tailor-made suit, useful for concretely modifying the organizational context, building awareness and making new ways of working possible.
For this lever to produce results, learning and new organizational conditions must evolve together: clear rules, accessible tools and sufficient time for testing allow you to put what you have learned into practice. The mapping therefore helps to orient both the training path and the attention to the conditions that make its application possible. At MADE we value experiential training precisely to bring the moment in which we learn closer to the moment in which we apply what we have learned.
Personally, I consider returning to work an essential part of the path started: it is there that doubts, obstacles and new questions emerge. This is why the comparison should not end with the last lesson. We need to gather feedback and check what has been applied. Reading before and after the training can help, but it is not enough on its own to encourage an economic return: that requires continuity and coherence with the project.
What are the most urgent and requested digital and AI skills in the manufacturing and industrial sector today that SMEs struggle most to find or develop?
If I had to indicate a skill to start from, I would say: knowing how to judge the scenario in front of us. With AI, convincing answers can be obtained quickly. Understanding whether they are also reliable requires attention and critical thinking. This means, first of all, knowing how to work with data: knowing its origin, understanding its limits, obtaining useful information from it. Management, data analysis and cybersecurity, moreover, emerge among the priorities also in the research of the Polytechnic.
Added to these basics are very practical skills: choosing when to use AI, formulating correctly contextualized requests, verifying results and protecting company information.
In manufacturing, knowledge of processes remains essential. In technical documentation, quality or maintenance, for example, it is that knowledge that allows you to ask the AI a good question and notice when an answer doesn’t add up. For this reason, training must involve not only computer scientists, but also those who know the factory, the product and the daily difficulties. It is from the encounter between technological skills and knowledge of processes that useful applications for the business can arise.
On the other hand, not all SMEs need to develop algorithms in-house, but they all need to build the skills needed to identify opportunities, talk to technology partners and integrate the right solutions into their work. The point is this: we need to empower people to drive digitalization and translate it into improvements for the company.
Then there is a skill that I feel is increasingly urgent to build in SMEs: the ability to collect and make transferable to new hires the knowledge of those who, due to age or professional choices, are about to leave the company. Generational change is decisive for the continuity and growth of a company, and here AI can have a fundamental role: helping to structure and make available the technical know-how that often remains only in the minds of the most expert people, so that those who join the company can access it without having to start from scratch.
If we were to make a (difficult, I realize) medium-long term forecast, how could corporate training change in the next 3-5 years thanks to the integration of AI-based analytical tools?
I believe that in the next three to five years we will have more opportunities to build training that meets everyone’s needs. AI will be able to help us recognize needs, adapt paths over time and offer opportunities to practice when needed, even outside the classroom.
However, when I think about what will really make a difference, I always come back to the human factor. Before the tools comes the mindset: the curiosity to understand, the willingness to try and also to admit “I don’t know this yet”. It is an attitude that is anything but obvious, especially when you have experience and are used to doing your job well. Learning something new sometimes means accepting that you feel like a beginner again.
