AI, companies ready but skills need to be reorganised

Even the most wonderful technologies do not change a company until they are adopted to change a process or create new products or services. Artificial intelligence is no exception; indeed, adoption is an integral part of the creative process. It is not a technology that is bought and used: it must be thoroughly understood, designed, experimented with, in order to be trusted.

"With artificial intelligence you innovate and accelerate business, you don't just file down a few costs," says Enza Truzzolillo, CEO of Lenovo Italia. "Value projects are designed in relation to the characteristics of each individual company. It follows that adoption is not an immediate process.

A window on this reality can be found in the results of the survey carried out by a Young Entrepreneurs Committee of Confindustria in cooperation with Lenovo. 621 entrepreneurs participated in the survey. Half of these companies are growing and 23 per cent have an export share of turnover of more than 30 per cent. A quarter of these companies have a turnover of more than EUR 10 million: all the others are smaller. Well: among these companies only 18.7 per cent use intelligence in a structured way, two thirds are in the study or test phase. Seventy-one per cent see it as an opportunity. Eighty-one per cent assess that Italian companies are lagging behind European companies. The central problem is seen in the preparation of personnel.

But behind these numbers Truzzolillo sees the variety of companies' stories. "Artificial intelligence is an accelerator of results. It serves to enhance the data and values of companies in order to create a shared knowledge on the basis of which innovative processes can be triggered'. It is first and foremost an organisational issue. "The first step is to define the company's distinctive know-how and separate it from other data. What feeds the core business should be used with artificial intelligences that are in house, with 'tailor-made' projects. What instead does not constitute the core business can be handled in the cloud with industrial models. It is an organisational problem'. And the research confirms this: 58 per cent of respondents believe that the most urgent skills to be developed relate to change management. A minority mentions technical skills.

"But because artificial intelligence is expensive, in terms of energy and technology, priorities have to be set. Lenovo operates with a hybrid approach, whereby it offers the technology and support for projects that partly involve data that does not have to leave the company and partly can use cloud and large models. Case examples? At the European Institute of Oncology, where data is very sensitive, the project was geared towards developing sophisticated diagnostic predictive models in the user's home to realise personalised services. At Corrado La Forgia's Vhit, which operates in automotive components, artificial intelligence has entered the production process to drastically reduce errors in products. And at The Edge Company, which deals with security at airports, a physical AI model has been developed that enables real-time analysis of all the data needed to predict the movements of flocks of birds that could make aircraft take-off difficult. These are examples of projects with systems that are protected against know-how leaks, suitable for creating new services and products or improving quality. Nothing to do with simple cost reductions. "Artificial intelligence is not the same for everyone. Like innovation, of which it becomes an essential component,' Truzzolillo concludes.