When it comes to innovation, large companies and SMEs often move at two different speeds. The former, thanks to substantial resources, dedicated teams, and a greater appetite for risk, are able to adopt new technological paradigms rapidly, effectively creating a gap with SMEs. The latter, however, are held back by tighter budgets, skills that are not always up to date, and a more cautious approach toward investments with uncertain returns.
This gap, already evident on several past occasions (e.g., with the cloud), reappears today with artificial intelligence. In this article, we will analyse the causes of this disparity and, above all, seek to understand how SMEs can overcome it to make the most of the competitive and efficiency advantages offered by AI.
Artificial intelligence in SMEs: few concrete projects
Artificial intelligence is gaining ground in the Italian corporate landscape, with a 58% increase in investments in 2024, totalling €1.2 billion. However, this momentum has not yet translated into widespread AI adoption among SMEs, which, despite representing the beating heart of our economic fabric, struggle to turn theoretical interest into concrete projects.
According to the AI Observatory of the Politecnico di Milano, only 7% of small enterprises and 15% of medium-sized ones have launched AI-based initiatives, primarily focused on operational efficiency and the optimization of production processes. These figures remain modest compared to the technology’s potential, signalling a gap between the desire to innovate and the ability to do so.
Confirming this trend is the survey “SMEs in the AI Era: diffusion, opportunities and prospects” by Confalpi and the Labour Consultants Foundation, which highlights strong interest in AI but still limited adoption. Indeed, 47.6% of SMEs declare themselves merely curious about the technology, while 29.1% show high interest. Yet only 11% have already developed AI-based solutions, rising to 29.7% if one also counts companies that, without concrete implementations, participate in pilot projects or training activities.
Why do SMEs struggle to adopt AI? The real causes, beyond the budget
Superficially, one tends to attribute SMEs’ lag to smaller financial resources compared to large enterprises. However, deeper analysis reveals numerous structural factors that explain this gap.
Immaturity in data management
The Politecnico di Milano Observatory maintains that many SMEs limit themselves to sporadic analyses, which, we would add, precludes the development of a widespread, structured data culture. AI relies precisely on the availability of well-organized and continuously updated data, but without consolidated processes for data collection, storage, and analysis, any AI project risks never reaching completion.
Lack of specialised technical skills
Citing the Confalpi research, 47.7% of companies identify the skills shortage as the main obstacle to AI diffusion. This problem has multiple causes: on one hand, access to specialised talent is complex, since SMEs struggle to compete with large companies in terms of attractiveness, career prospects, and financial offers; on the other hand, in-house AI training remains limited, leaving many companies without the skills needed to integrate the technologies effectively.
Resistance to Innovation
AI adoption is not only a technological issue but also a cultural one. Many workers fear that these technologies could negatively affect their roles and tasks, creating a psychological barrier to innovation. A change-oriented mindset and effective change management are essential to overcome these resistances, creating fertile ground for continuous corporate evolution.
Inadequate infrastructure
AI adoption requires systems capable of collecting, storing, and processing large amounts of data. Some SMEs still rely on outdated and poorly scalable infrastructures that do not meet the needs of modern AI systems. This issue is directly linked to investment: without adequate IT infrastructure, any AI project becomes difficult to realise. For SMEs, the cloud is more than an adequate solution.
Regulatory Uncertainty
Companies find themselves navigating a constantly evolving regulatory landscape, which generates uncertainty. According to Confalpi’s research, 12% of SMEs cite regulatory uncertainty as a brake on AI adoption, a factor that cannot be ignored. In Europe, where regulations are often complex and ever-changing, between local rules and international laws, it is understandable that companies may have doubts about what will be permitted and what adjustments will be necessary. In this scenario, the new AI Act will become a reference point, as it could bring simplification and certainty, offering SMEs clear guidance.
AI and SMEs: the implementation roadmap
If it is true that over 50% of SMEs intend to invest in AI over the next three years, it is essential to understand how to do so effectively. AI adoption cannot be viewed as a process centered solely on technology; rather, it must be a systemic approach involving the entire company—its culture, internal assets, skills, and organisational structure. We have identified a possible roadmap.
Training and Reskilling
Considering the obstacles to widespread AI adoption, the first step is investment in (digital) training. This can be structured on two distinct yet complementary levels:
- Technical training aimed at reskilling key professionals within the company. These figures, often hard to recruit externally, are fundamental for managing the development, integration, and evolution of AI solutions;
- Digital training for the entire workforce, to foster a genuine data culture that becomes an integral part of every business function. Building a shared knowledge base not only increases the effectiveness of adopted technologies but also develops an innovation-oriented mindset that supports adaptation and continuous evolution.
Engaging expert and trusted consultants
A systemic transformation journey must be guided by experienced professionals. It is crucial to rely on those who not only propose and implement technological solutions but also offer strategic vision and support the partner throughout the adoption process, including training and change management.
Developing pilot projects on specific use cases
Before embarking on large-scale projects, it is advisable to identify specific use cases where AI can deliver concrete, measurable value. In the past, some companies have attempted overly ambitious approaches—such as replacing operators with chatbots in critical areas like customer service—with disappointing results and reduced customer satisfaction. It is therefore more advantageous to start with targeted projects, focusing on clear objectives of operational efficiency and productivity improvement. This approach enables companies to explore AI’s potential in a controlled manner, reducing risks and obtaining immediate feedback.
Large-scale implementation and monitoring
The success of initial projects opens the door to creating a centralized data platform capable of feeding all current and future AI implementations. This platform not only ensures a coherent data flow but also guarantees centralized governance, compliance with current regulations, and adequate data protection.
At the same time, it is essential to monitor each project’s performance, not only from a technical standpoint but also in terms of business impact and cultural evolution. How is the company reacting to AI introduction? Are employees using data and solutions correctly in their daily processes? Is security being maintained?
Continuous evolution
AI adoption must be an ongoing, ever-evolving process. SMEs must be ready to refine their solutions and expand them to new use cases, always with an integrated and strategic approach. Only in this way can AI become a continuously growing resource, capable of generating increasing value for the business. It is essential to remember that we have only seen the tip of the AI iceberg so far, and its potential is destined to grow exponentially alongside underlying technological developments.