Innovation, Ai enters the company but in small doses and with caution

Discussed but how widely adopted and considered decisive? The Politecnico di Milano uses a reference to a literary imagery, that of the play Waiting for Godot, to describe the relationship that many companies have today with artificial intelligence. In fact, the quotation from Samuel Beckett's masterpiece opens the just-released report entitled 'Ai4Innovation - Innovation & Ai in Italian companies: Gen-Ai & Agentic-Ai between awareness, prudence and action'. "The data show - we read - that despite the technological acceleration of the last two years, the diffusion of Ai in business innovation processes still remains limited. Many organisations observe, wait, evaluate - but act very cautiously'.

The survey was conducted between February and March 2026 among 85 business organisations distributed over four size brackets by turnover. The largest share is accounted for by medium-sized companies (turnover between EUR 10 and 50 million) and large companies (between EUR 50 million and 1 billion), each accounting for 27% of the total, followed by small companies (turnover less than EUR 10 million), which make up 32% of the sample. Very large companies, with turnover above EUR 1 billion, account for 14% of the respondents.

Starting with the first question, which asked people to indicate their organisation's level of maturity in adopting GenAi and Agentic Ai, the answers were aggregated by drawing three profiles. Each of these reflects progressive stages of AI integration into business processes.

Organisations are then Ai Starters (25%), i.e. they still approach Ai in a sporadic and unstructured manner, more driven by an individual initiative than by a defined strategy. Or they are Ai Experimenters (49%), i.e. they have started pilot projects on specific situations, in this case it is a partial and not yet systematic adoption: 'Ai supports some activities, but it has not yet entered the operational routine'. The last profile, the Ai Scalers (26%), encompasses organisations that have passed the experimental phase, 'whereby Ai is adopted - the researchers explain - on several use cases in a structured manner, with defined governance, clear ownership and monitoring metrics'.

But what is happening in terms of training and skills? Here, the survey investigated whether and to what extent organisations perceive the need to develop new skills within innovation teams, and with what levers they are trying to fill the identified gaps.