AI Statistics for Italian SMEs 2026: Only 15.7% Use AI — Double 2024's 7.7% (Istat & Polimi Data)

Only 15.7% of Italian SMEs Use at Least One AI Technology — but Adoption Has Doubled in a Single Year, from 7.7% in 2024 (Istat, December 2025)

Only 15.7% of Italian SMEs use at least one AI technology — but adoption has doubled in a single year, from 7.7% in 2024, according to the "Imprese e Ict – Anno 2025" report published by Istat on 15 December 2025. It is the single most important statistic released this year on artificial intelligence in Italian small and medium enterprises, and it tells two stories at once: roughly 85% of SMEs are still on the sidelines, but the minority that has moved is growing at the fastest rate ever recorded. This page collects the 2025-2026 statistics on AI and automation in Italian SMEs with the primary source linked next to every number — Istat, the three observatories of Politecnico di Milano, Webidoo's SME-AIMIX index, and documented company cases — so you can quote them without reconstructing where they came from.

The National Numbers for 2026: What Istat and Politecnico di Milano Report

The most solid snapshot of AI adoption in Italian companies is the Istat "Imprese e Ict – Anno 2025" survey (press release of 15 December 2025, data referring to 2025). The picture at a glance:

The time series makes the acceleration even clearer: across all firms with at least 10 employees, adoption went from 5.0% in 2023 to 8.2% in 2024 to 16.4% in 2025 — a doubling per year, two years in a row. And yet 83.6% of Italian companies still use no artificial intelligence technology whatsoever.

The market context pushes in the same direction: according to the Osservatorio Artificial Intelligence at Politecnico di Milano, the Italian artificial intelligence market reached €1.8 billion in 2025, growing 50% in a single year. Most of that spending currently comes from large enterprises, but it is precisely this critical mass that has made the tools — language models, automation platforms, cloud services — more mature, cheaper, and therefore finally within reach of an SME with a four- or five-figure budget, not a seven-figure one.

Politecnico di Milano adds the investment and skills dimension. According to the Osservatorio Innovazione Digitale nelle PMI (press release of 21 May 2026), 76% of Italian SMEs have neither invested nor plan to invest in artificial intelligence, only 7% have started structured AI training programs for their staff, and 47% have done no research and development in the last three years. Heavy numbers, considering that Italian SMEs are more than 240,000 firms generating over 40% of national revenue.

On the process automation front, the Osservatorio Intelligent Business Process Automation (press release of 4 March 2026) measures adoption of 30% — up 7 points on 2024 — for AI-powered "intelligent" automation — but note: that figure refers to large enterprises. Among SMEs, process automation adoption stops at 15%, while 22% of SME working time is still absorbed by repetitive tasks. Agentic automation trials stand at 8% in large firms and just 2% in SMEs.

Why do so many companies stay put? Istat is blunt: the lack of adequate skills is the number-one obstacle, cited by 58.6% of firms that evaluated AI investments and then did not go through with them — nearly 60%. It is not a technology problem and, in many cases, not a budget problem: it is a problem of knowing where to start.

The Territorial Divide: North-West Leading, South and Islands Halfway There

AI adoption among SMEs is not uniform across the country. The SME-AIMIX 2026 index by Webidoo Insight Lab (via Adnkronos) measures this divide four ways:

Istat's own territorial figure — covering all firms with 10+ employees, not SMEs only — confirms the direction: companies in the North-West show the sharpest growth, going from 8.9% in 2024 to 19.3% in 2025. Between an average North-Western firm and a Southern one there is a gap of more than 7 percentage points: companies in the South start further back, but that is exactly why early adoption there is worth more in local competitive advantage.

SMEs vs Large Enterprises: A Gap That Keeps Widening

The paradox of 2025 is that adoption doubles everywhere, yet the absolute gap grows. Large enterprises went from 32.5% to 53.1%: +20.6 points in one year, against the 8 points gained by SMEs. Today a large Italian company is more than three times as likely to use AI as an SME (53.1% vs 15.7%).

The same pattern holds for process automation: 62% of large enterprises use at least one process automation technology (+10 points on 2024), against 15% of SMEs. Meanwhile, nearly a quarter of SME working time — 22% — goes into repetitive tasks that large companies are already delegating to machines. To see how this plays out in one concrete operational area, we analysed warehouse management automation for SMEs.

The good news, if you run an SME, is that documented cases show proportionally higher returns precisely in small organisations, where a single automated process can free up a significant share of total work. The three cases below — with the caveats stated — prove it.

Studio Epica: 12,000 Invoices Automated and a Claimed 1,420% ROI

Studio Epica is an Italian accounting firm that automated invoice bookkeeping with an RPA bot built by Digital Automations. According to the case study published by the vendor — these are figures reported by the agency that delivered the project, not independently verified — the results are as follows:

Even discounting the optimism typical of vendor case studies, the order of magnitude is credible: invoice bookkeeping is the quintessential repetitive process — high volume, stable rules, zero creativity — and exactly the kind of activity where automation pays off fastest. We collected more examples with similar numbers in our review of AI and automation case studies from Italian SMEs with documented ROI.

Rold: The Italian SME That Entered the World Economic Forum's Global Lighthouse Network

Rold, a company from Cerro Maggiore (Milan) with roughly 250 employees making components for household appliances, is the historic proof that an Italian SME can compete in industrial AI at the highest level: it was the only Italian SME admitted to the World Economic Forum's Global Lighthouse Network, the club of the world's most advanced factories. With its SmartFab platform, developed in-house with Samsung, Rold achieved an 11% improvement in OEE (overall equipment effectiveness) that translated into a 7-8% revenue increase over 2016-2017.

The numbers are almost a decade old — read them as the starting point of intelligent Italian manufacturing, not a current benchmark — but the lesson stands: you do not need to be a multinational to do Industry 4.0 seriously. For the updated sector picture, see our analysis of AI in Italian SME manufacturing.

SACE: The Large-Enterprise Benchmark (to See How Far It Can Go)

SACE is not an SME — it is an insurance-financial group — and we include it only as a top-end reference point. According to figures published by Microsoft Italy (again: a vendor source), after 11 months of using Microsoft 365 Copilot the company recorded a 25% productivity increase, 10 hours saved per user per month, email management time cut in half, and a 97% usage rate. SACE's 10 hours per user per month and the 5.2 hours per week from the OpenAI research tell the same story from two angles: the first benefit of AI in a company is time handed back to people.

What This Means for Your SME

Three practical takeaways emerge from this data:

• The competitive-advantage window is now. With roughly 84% of SMEs still on the sidelines and adoption doubling every year, whoever starts in 2026 is among the first in their sector and territory; whoever waits until 2028 will be chasing.
• Start from the 22% of repetitive time, not from the technology. The Polimi figure on time absorbed by repetitive tasks is the treasure map: invoices, quotes, data entry, standard replies. These are the fastest-ROI processes, as the Studio Epica case shows.
• The real bottleneck is skills — and it can be worked around. 58.6% of companies stall for lack of skills and only 7% train their staff. You do not need to hire a data scientist: you need a partner who identifies the right processes and transfers the know-how while automating them.

One last note for anyone writing, quoting, or deciding on the basis of this data: these numbers change fast. Adoption doubled in twelve months, and the next waves of the Istat and Politecnico observatory surveys are expected between late 2026 and spring 2027. If you reuse these statistics in an article or a presentation, always state the reference year of the figure (2025) and link the primary source: it is the difference between a credible analysis and a number out of context.

Sources & References

• Istat — "Imprese e Ict, Anno 2025", press release, 15 December 2025 (PDF with annex tables)
• Osservatorio Innovazione Digitale nelle PMI, Politecnico di Milano — press release, 21 May 2026
• Osservatorio Intelligent Business Process Automation, Politecnico di Milano — press release, 4 March 2026
• Osservatorio Artificial Intelligence, Politecnico di Milano — Italian AI market 2025
• SME-AIMIX 2026, Webidoo Insight Lab — via Adnkronos
• OpenAI research on AI time savings in Italian SMEs — via AI4Business
• Digital Automations — Studio Epica case study (vendor-reported)
• Innovation Post — Rold in the WEF Global Lighthouse Network
• Microsoft Italy — SACE results with Microsoft 365 Copilot (vendor-reported)