en.Wedoany.com Reported - Dedagroup has launched the ESS-AI framework, aiming to transform artificial intelligence (AI) into a stable lever for productivity and governance for Italian businesses and the public sector, marking the entry of industrial AI into a responsible phase. The group believes that merely experimenting with isolated models or prototypes is no longer sufficient; the decisive step lies in the governance of the entire supply chain, where data, infrastructure, models, applications, and skills must work in synergy.
ESS-AI, which stands for European Secure & Sovereign AI, is positioned by Dedagroup as a full-stack ecosystem encompassing technology platforms, advanced language models, supercomputing, sovereign solution orchestration, and vertical domain expertise, with the entire system located in Italy. This initiative reflects a shift in AI adoption: the focus moves from the ability to use powerful models to enabling data circulation while maintaining governance, security, and accountability. Dedagroup defines this balance as "risk sovereignty," where the emphasis shifts from abstract data ownership to the ability to decide how, when, and with whom to share data.
The goal of ESS-AI is to guide enterprises and institutions toward an AI-empowered business model, operating within processes governed by conscious humans, ensuring that automation is subordinate to human choice, trust standards, and the context of use. Marco Podini, CEO and Executive Chairman of Dedagroup, emphasized that each organization's ability to define its own risk sovereignty based on information sensitivity, usage context, and the trust levels of all parties is the most significant game in the digital age.
The framework is built around five integrated pillars. First, activating enterprise knowledge by making it queryable through Corporate Large Language Models (LLMs) and specialized Small Language Models (SLMs). Second, developing precise AI solutions for prediction, optimization, and risk management. Third, building sovereign and secure AI infrastructure designed for critical workloads and data protection. Fourth, constructing vertical platforms by combining proprietary models with domain expertise. Fifth, focusing on AI adoption, skill development, and informed use to drive organizational change.
At the core of this supply chain is Istella, which Dedagroup describes as Italy's largest knowledge graph, containing over 6 billion web page indexes (half of which are in Italian) along with billions of social interactions and multimedia content. Based on this, enterprise LLMs and SLMs trained on Italian data have been developed, adaptable to client application domains. The precision with which the knowledge graph connects, interprets, and uses data is seen as a competitive advantage, helping to strengthen European industrial autonomy.
On the infrastructure front, Dedagroup mentions Intacture, an underground data center located within an active mine in Trentino, of which the group is a private founding member. This asset offers advantages in computing power, physical security, data sovereignty, and integration with the European ecosystem. The emphasis on supercomputing underscores that having controlled, localized infrastructure capable of integrating with European networks is an industrial issue. Sovereignty here is defined as verifiable control, allowing enterprises to use AI without relinquishing data governance.
Application development is handled by Deda's AI Innovation Center, focusing on data valorization, proprietary algorithms, agentic AI models, and vertical solutions. Agentic AI can directly act on operational processes in a traceable and compliant manner. Dedagroup mentions solutions already running in regulated verticals, including Unica Wealth Solutions for banking, EMIRA for the energy sector, the SaaS platform Refer for document management, and the composable ERP platform Stealth Cosmica for the fashion industry.
Dedagroup has integrated structured education, retraining, and upskilling pathways into its offerings, noting that the group's AI Competence Center has over 200 professionals. Skills are considered part of the infrastructure, as AI only generates value when users understand its limitations and know how to integrate it into workflows.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
