AI: Italy Moves from Experimentation to Operations. SAP Study

How much do Italian companies spend on generative and agentic AI? According to a study by SAP and Oxford Economics, the average is $18.4 million, a figure considerably lower than the $28 million spent on average globally. However, the analysis also forecasts a 45% growth in investments over the next two years. These investments are driven by the ROI from adopting these technologies, which currently stands at around 20% for companies and is expected to rise to 38% over the next two years.

Investments in AI yield concrete results. Italian companies are increasingly mature in AI practices. The figure of €18.4 million invested in AI may seem quite high, but it's important to consider the sample: the study was conducted by interviewing 2,600 executives from large companies or mid-caps in 13 countries. We are not talking about the typical SMEs, but about companies of a certain size. Mid-caps, for clarity, are businesses that generate up to $150 million in revenue. In Italy, the average revenue for companies in this bracket falls between €30 million and €50 million, many of which operate in the manufacturing sector.

The data reveals an interesting fact: the experimentation phase is over, and companies are beginning to successfully use AI. The next step, according to SAP, will be to invest more in agentic AI.

How Italian Companies Use AI

The research highlights that as of now, about a quarter of business activities (26%) are already supported by AI, a percentage expected to reach 44% within two years. However, it should be noted that approaches vary widely: while 20% have made strategic investments, 36% of companies still adopt a fragmented approach, which hinders fully exploiting the technology's potential.

The reasons are various. Only 45% of the companies examined have a dedicated individual responsible for integrating AI into the business, who should define KPIs, prepare training courses, and raise awareness of the benefits and risks of this technology. Yet, despite these challenges, three out of four companies (73%) are satisfied with the ROI achieved, being aware that there is room for improvement.

Expectations are also supported by agentic AI: 75% of the companies surveyed believe it can have a moderate to high impact on organizational transformation. However, the level of maturity remains low. Just 1% of the sample considers itself fully prepared to adopt it, while most companies indicate they are only partially ready or not yet prepared at all.

Challenges to AI Adoption

The quality of data remains the primary obstacle to AI adoption. Although 67% of Italian companies consider themselves prepared in this regard, 74% report issues related to the reliability of the available information. The consequences directly affect operational processes: 77% of companies have had to deal with rework, delays, or backlogs due to poor quality of outputs generated by AI systems. A second area of concern revolves around skills. 80% of companies do not believe that professional training programs keep pace with the evolution of AI tools.

Almost no respondents think that technology will not have consequences for workforce planning. To complicate matters, Shadow AI also contributes to the picture: 67% of the sample reports that unauthorized use of AI tools occurs at least occasionally within the organization.

Finally, the issue of governance remains open. Only one in ten companies considers itself fully prepared to manage AI effectively: 11% in terms of skills and 10% regarding processes, rules, and organizational frameworks.

"In the last year, Italy has moved from the experimentation phase of AI to operational status, and this is beginning to yield concrete results. However, there is still much work to be done. An AI lacking context, whether in processes, data, or governance, at best generates activities without results, at worst creates new risks," explains Carla Masperi, CEO of SAP Italy. "The next step to derive value from the new technology will be to integrate data, processes, and AI. Italian companies must understand that AI often provides value that is more difficult to measure than expected and entails risks that evolve at a very fast pace for most governance systems to keep up. For instance, AI governance is becoming the most significant challenge that many companies are still unaware they have."