AI adoption in Italy lags EU, boosts long-term productivity

The share of Italian firms with at least 20 employees using AI tools reached 32 percent in early 2026, up from 27 percent in 2025, according to a Banca d'Italia survey.

However, intensive use remains limited at 5 percent.

Italy's adoption rate is four points below the EU average and nearly ten points lower than Germany's.

While AI is primarily used to optimize existing processes rather than develop new products, long-term simulations indicate potential productivity gains of 0.2-1.1 percentage points annually over ten years, depending on the speed and depth of adoption.

Short-term effects on firm-level productivity are not yet significant, consistent with literature suggesting aggregate benefits materialize after substantial organizational adjustments.

The paper identifies key barriers to AI adoption, including difficulties for less structured firms in evaluating gains, identifying suitable uses, and securing capital and skills.

Externalities and network effects mean individual firms may not capture all benefits, potentially leading to suboptimal adoption levels.

Public intervention is justified to remove informational and organizational obstacles, develop skills, and narrow the gap between private and social returns.

This is particularly relevant for AI, given its novelty and firms' limited understanding of its potential.

Fragmented demand from small firms can also hinder the supply of specialized AI solutions, creating a vicious cycle that public policies can break.