There's a convenient narrative that artificial intelligence is something for big corporations: huge budgets, data scientists on the payroll, dedicated infrastructure. It's a convenient narrative because it lets those who never start off the hook. But the numbers tell a different story: the problem for SMEs isn't that AI isn't for them, it's that almost no one has made it accessible to them in a concrete way.
The gap, in three numbers
The AI market in Italy is moving fast. Adoption by companies, much less so, and in deeply uneven ways depending on size. Looking at who has adopted AI in a structured way (not the occasional use of a chatbot, but integration into processes):
Seventy-one percent versus seven. This isn't a subtle difference: it's a fracture. And it's not because SMEs don't have the same problems, slow processes, repetitive tasks, knowledge that lives in the heads of a few people, but because until now they've been poorly served. Source: Osservatori Politecnico di Milano, Banca d'Italia.
Why SMEs stay behind
When you ask Italian companies why they don't use AI, the answers almost always converge on two concrete obstacles, the very same two points, by the way, on which the whole game is played.
1 · The skills they lack
Integrating AI isn't about writing a prompt. It requires understanding which processes are worth touching, redesigning them, and above all putting the team in a position to use them every day. Large companies hire people who know how to do it. An SME doesn't, and this is where dependence on outside vendors begins, which is exactly what holds it back.
2 · Concerns about data
The second barrier is confidentiality: "where does our data end up?". It's a legitimate concern, not an alibi. And it's solvable, with governable architectures, compliance-by-design and clear rules about what goes in and what doesn't, but it has to be addressed, not dismissed.
The question isn't whether SMEs need AI. They do, often more than large companies, because every hour freed up matters more. The question is making it adoptable: skills transferred to the team and data under control.
The hidden advantage of SMEs
There's a reversal that almost no one talks about: on a well-run AI journey, an SME starts out at an advantage over a large company. Fewer decision-making layers, processes that one person knows in full, the ability to see a result in weeks rather than quarters. Where the large company has to align committees, the SME can decide on Monday and start on Wednesday.
The 71% versus 7% gap, read this way, isn't a sentence: it's the space in which an SME that moves now gets there before its competitors.
Who it really works for
Being honest also means saying who it's not the right time for. AI in a company pays off when there's a concrete operational problem to solve, not when you simply want to "do AI" so as not to fall behind. The profile where the impact is greatest:
| Dimension | Ideal profile |
|---|---|
| Revenue | Over €2.5 million, with operating costs and processes to optimize |
| Headcount | From 20 to 250 employees: large enough to have processes, agile enough to move |
| Attitude | Curious but disoriented: feels the pressure of the market, doesn't know where to start |
| Frustration | Has already seen consulting with no results. Wants deliverables, not slides |
By size and by the nature of their processes, the sectors where we see the greatest return are manufacturing, pharmaceuticals and healthcare, financial services, retail and e-commerce, distribution and logistics, food & beverage, energy and utilities, professional services.
Where to start
The first step isn't buying a tool: it's understanding where you stand. A serious AI Readiness Assessment identifies the few processes where AI makes the biggest difference, estimates the impact and defines the KPIs before spending a single euro on implementation. From there, a journey that transfers skills to the internal team and keeps data under control, so that in the end you don't depend on anyone.
The 7% is a starting point, not a destiny. The SMEs that close the gap now don't do it with bigger budgets: they do it with a method.