The survey on artificial intelligence in the large-scale retail sector, conducted in collaboration with the tech companies Aton and GTN, E.N.I.A. (the Italian Foundation for Artificial Intelligence), and Fòrema, the training body of the Confindustria Veneto Est system, reveals that companies are experimenting with innovative solutions, but only one in seventy has reached the industrial scaling phase. Governance and compliance with the AI Act are now essential priorities. Giorgio De Nardi, CEO of Aton – The future of retail will depend on the ability to capitalize on this extraordinary opportunity to support daily work and decision-making, while rigorously measuring its real-world impact.
The Italian large-scale retail sector no longer views artificial intelligence as a futuristic concept, but as an imminent operational necessity. However, the path from theory to industrial practice is still a long one.
The “AI in Large-Scale Retail” survey, conducted by Fòrema and E.N.I.A. and commissioned by the tech companies Aton and GTN among a sample of 70 leading companies in the sector (there are just over 100 in Italy, 46% of which have revenues exceeding one billion euros), paints a picture where enthusiasm for the technology’s potential coexists with marked organizational caution. The survey captures a precise snapshot: we are still in the midst of the exploratory testing phase. This observation also applies to companies based in the Veneto region: 17 were interviewed, 5 of which are headquartered in the Padua area.
The study was commissioned by the tech companies Aton, headquartered in Villorba in the Treviso area, and GTN, an IT company based in Tavagnacco in the province of Udine, and presented in Padua at Le Village during the “Retail Tech” event. Aton and GTN, which specialize in digital solutions for industry and distribution, both focus on the digitization of commercial activities and traceability. Their collaboration has created a technology cluster with 385 direct employees and 33 million in revenue. The partners supporting the initiative are prominent: E.N.I.A., the National Foundation for Artificial Intelligence, Fòrema (which collaborated on drafting the survey), Largo Consumo (media partner), Zebra Technologies, Datalogic, and Soti (tech partners).
The most striking finding concerns the digital maturity curve. While 68% of respondents (equivalent to 48 companies) report being in the exploration or active experimentation phases, only a single company (representing approximately 1.4% of the sample) reports having begun “scaling”—that is, adoption on an industrial scale. Conversely, 19% of companies (13 businesses) have not yet initiated any activities. This imbalance represents the true bottleneck of modern retail. It is not just a matter of making a new technology work, but of making it replicable and sustainable within complex logistics networks and across hundreds of physical retail locations.
Where artificial intelligence comes into play, it does so by prioritizing pragmatism over spectacle, focusing directly on internal efficiency. The most common use cases are led by corporate “knowledge bases,” adopted by 38 companies, followed by out-of-stock monitoring, with 28 applications. Automated document reconciliation (DDT) and customer service chatbots each account for 23 implementations, while price and promotion management involves 28 companies. Finally, energy consumption optimization accounts for 21 applications, demonstrating a focus primarily on reducing operating costs.
Paradoxically, the main obstacle to widespread adoption is not related to technological or financial limitations. The real barriers identified by managers are organizational in nature. There is a lack of comprehensive governance that would allow for the implementation of AI in a streamlined and secure manner. The shortage of specialized internal expertise is cited by as many as 41 respondents, closely followed by cultural resistance to change and mistrust of AI (35 responses). The road is also uphill on the regulatory front: only 11 companies have formalized an internal AI governance policy, 17 are in the process of defining one, while as many as 42 organizations have not yet addressed the issue of compliance with the European AI Act. Uncertainty regarding the economic return is also a factor: for 29 companies, the real benefits of this innovation remain “to be quantified.”
COMMENTS
The strong push toward logistics and back-office efficiency documented by the survey shows that large-scale retailers believe in the concrete and reliable value of AI, – says Giorgio De Nardi, CEO of Aton. – However, the true test for modern retail remains the physical store, where the integration of artificial intelligence, information systems, and on-the-ground staff is still complex. To move beyond the pilot project phase and make a real leap to industrial scale, a systemic vision is needed, one capable of translating data into immediate operational decisions that improve speed and accuracy.
The data confirms that the real obstacle to the adoption of artificial intelligence in large-scale retail lies not in the technology, but in skills and organizational culture – says Matteo Sinigaglia, general manager of Fòrema – The fact that 41 out of 70 companies complain about a lack of in-house experts and that the vast majority have not yet defined a governance policy shows that the digital transition risks stalling. Funding pilot projects is not enough; we need an extraordinary plan for data literacy and managerial training. Only by overcoming cultural reservations will Italian retail succeed in transforming algorithms into industrial productivity.
The survey clearly shows that the real challenge for large-scale retailers is no longer technological, but regulatory and governance-related,– says Paola Geretti, CEO of GTN – The fact that as many as 42 out of 70 companies have not yet addressed the issue of compliance with the European AI Act represents an extremely high strategic and operational risk. Adopting AI without a formal policy exposes companies to costly legal and ethical vulnerabilities. To unlock investments and scale projects safely, the large-scale retail sector must urgently establish clear governance models capable of balancing the push toward automation with full accountability for data.
The fact that just 1% of large-scale retail companies have reached the industrial scaling phase, compared to 68% that are still in the exploratory phase, demonstrates that we are facing a delicate cultural and regulatory shock—even before a technological one – comments Valeria Lazzaroli, president of the Enia Foundation – We note with concern that as many as 42 major companies in the sector have not yet addressed the issue of compliance with the European AI Act and that only 11 have formalized an internal governance policy. Artificial intelligence cannot be introduced in silos or treated as a mere test of operational efficiency to reduce costs. Without a solid regulatory framework, without a widespread data culture, and, above all, without massive investments in internal skills to bridge the organizational gap, AI risks generating mistrust rather than value.
