AI in Italian Companies in 2026: Why Adoption Is Growing but Transformation Isn’t

40% of Italian companies have adopted artificial intelligence, compared to 30% in 2025: a 33% year‑over‑year increase, accelerating compared to last year’s growth rate. Yet, according to the “Unlocking Italy’s AI Potential” report, commissioned by Amazon Web Services (AWS) and independently conducted by the research firm Strand Partners on 1,000 Italian business leaders, the share of companies using AI to launch new products or services has dropped from 40% in 2025 to 34% this year. More companies adopt AI, fewer use it to innovate. It’s a paradox worth understanding.

The three levels of AI maturity in Italian companies

The report divides Italian companies into three levels of AI adoption maturity:

Basic level (58%): most companies use public chatbots and off‑the‑shelf AI tools, focusing on operational efficiency and incremental improvements rather than innovation or transformation.
Intermediate level (24%): AI has been integrated into multiple business functions, with measurable improvements in efficiency and customer experience.
Advanced level (13%): companies use more advanced AI systems, combining multiple models, autonomous or agentic AI. In this case, AI is used at a transformative level, enabling the development of innovative products and services and access to new markets. The percentage has increased from 11% last year.

The critical figure is that nearly six out of ten companies stop at the basic level. They use AI as they would any operational software: to do faster what they were already doing. Not to change how they work, what they offer, or how they compete.

The European figure makes the gap even clearer: the European average for the advanced level is 22%, while Italy is at 13%. Progress is slow—only two percentage points more than last year. At the current pace, according to the report’s projections, it would take until 2034 for even half of Italian companies to reach this level. It’s worth noting that SMEs, which make up the majority of the Italian economy and drive national averages, are even further behind: only 11% of those that have adopted AI have reached the advanced level.

What is slowing down transformative AI adoption in Italy

According to the report, the reasons are not random—three main structural obstacles hinder the leap toward advanced AI use.

1. Skill gap

The lack of skills is the most widespread barrier: 48% of companies cite it as the main obstacle to expanding AI, and 43% report that their internal staff is not yet adequately trained. The self‑assessment picture is telling: only 18% claim to have solid AI skills, while 52% acknowledge they need to improve, and 26% say they are still at the beginning of the journey. Perhaps the most revealing figure is another: 79% expect AI skills to become critical in their sector within the next five years. Awareness exists. Urgency does not.

The gap concerns not only the present. Looking at the technical roles considered most important in the next five years, companies point to software developers (46%), solutions architects (38%), and data governance or privacy managers (34%). Three roles that require long training paths and investments today, not tomorrow.

2. Lack of dedicated budget

Four out of ten companies do not have a specific AI budget. It’s not just about insufficient resources: 26% directly cite the lack of internal funds as an obstacle, but another 22% get stuck even earlier due to the absence of a clear ROI assessment. Without a stable budget line, projects rarely move beyond the pilot phase. Yet 86% of these same companies expect AI to weigh increasingly on IT spending over the next three years, with estimates placing it at 18% of the total budget. The problem is not vision—it’s the gap between intentions and concrete decisions.

The picture becomes more complex when considering the role of the public sector as a potential accelerator. 54% of companies consider government support, grants, and tax incentives crucial or very important in AI adoption decisions. But 32% cite the slowness and complexity of public procurement as a real obstacle to scaling their solutions. A more fluid public demand could unlock private investments that currently remain stuck.

3. Weight of regulatory compliance

A third of Italian companies’ tech spending—34%, up from 30% last year—already goes to regulatory compliance. And the trend shows no sign of reversing: 76% of companies say compliance costs have increased over the past three years, and 72% expect them to rise further. In this context, the progressive entry into force of the European AI Act, which introduces specific obligations for high‑risk AI systems, adds another layer of complexity. At the European level, the share of tech budget absorbed by compliance already reaches 42%—a sign of how heavily the regulatory framework weighs on investment decisions.

Agentic AI in Italy: what it is and how many companies are actually adopting it

The relevant evolution is agentic AI: systems capable of achieving specific goals with limited human supervision, organized into AI agents that mimic human decision‑making to solve problems in real time. Unlike traditional AI models, which operate within predefined constraints and require human intervention at every step, agentic AI demonstrates autonomy, goal‑oriented behavior, and adaptability. In practice, it doesn’t just respond to an instruction: it plans, reasons across multiple steps, calls external tools, and acts. An agentic system can, for example, analyze production data from a plant, identify an anomaly, trigger a maintenance procedure, and update the management system—without anyone issuing each individual command.

This is the direction of current technological development, but the Italian market is still far from this maturity. Only 18% of companies say they have heard of agentic AI; of these, just 2% have fully implemented it and 8% are experimenting. The gap with startups is stark: 74% of Italian startups say they are ready to adopt next‑generation AI technologies, compared to only 15% of companies overall. The remaining 40% of companies say they are slightly ready or not ready at all.

The issue is not distrust: once the technology is explained, 55% of companies say they are willing to use it. Those that have already adopted it report measurable results: 38% report faster decision‑making processes, 30% greater operational efficiency, and 20% improved scalability of operations. The gap is not interest—it’s exposure, competence, and experimentation capacity. Companies that have never heard of agentic AI cannot evaluate its adoption. Those that know it but do not implement it often lack the technical skills to design it and the budget to experiment in a structured way. It’s the same loop that blocks AI transformation in general: without continuous investment in training and infrastructure, even the most promising technologies remain theoretical.

Italy risks losing its most innovative companies

There is a signal that deserves separate attention, because it concerns the future more than the present. Italian startups are among the most advanced in adopting next‑generation AI, but 34% say they are seriously considering leaving Europe to accelerate their growth. Only 42% rule out this possibility; 24% are still undecided.

The reasons are consistent and revealing:

greater availability of funding (52%);
better international expansion capacity (48%); more direct access to global markets (44%); lower operating costs (42%);
stronger support ecosystem—incubators, accelerators, partnerships (40%);
more predictable regulation (38%).

This is not an escape from an Italy perceived as inadequate in absolute terms: it is founders following the path of least friction to the market. And in the AI era, where speed is a structural competitive advantage, every month counts.

The good news is that the exodus is not inevitable. Startups themselves clearly indicate what would make them stay: strong access to European customers and markets (63%), availability of venture and growth capital (50%), solid ties with local ecosystems—partners, universities, accelerators (46%)—and a stable, predictable political environment (44%). These are conditions that depend on system‑level choices—industrial policy, regulatory simplification, ecosystem investment—more than on individual companies.

Why investing deeply in AI is worth the effort

Organizations that have reached the advanced level share several characteristics. 62% of Italian companies already identify AI as a primary or high strategic priority, and 58% consider it a critical or important element of their overall strategy. But declared priority is not enough unless supported by continuous investment, internal skills, and a clear choice about where AI should create concrete value.

These companies have moved beyond the logic of isolated pilot projects: AI is integrated into core systems (ERP, production management, CRM), supported by structured, high‑quality data, clear governance, and trained personnel to manage it. It’s not about having more tools—it’s about having a precise strategy for where AI should make an impact and the organizational capacity to sustain it over time.

The shift from basic to advanced level also requires technical partners capable of translating AI potential into applicable implementations. The same companies interviewed identify the most important factors for supporting their growth: private investment and venture capital (55%), access to the most advanced AI tools (52%), and qualified AI talent (48%).

The data confirms that the leap is worth the effort. 72% of companies using AI in a structured way say their innovation cycles have accelerated over the past two years. Among startups, this figure rises to 88%. The challenge, as the report itself summarizes, is not increasing the number of companies using AI, but ensuring that more companies implement it deeply enough to truly innovate.

Sources: