Expert.ai and UCIMA Launch the First Italian LLM for Manufacturing
A strategic milestone for national industrial innovation and the practical adoption of AI in production processes achieved through funding from the PNRR on behalf of the MADE Competence Center 4.0 OR MADE 4.0 Competence Center – Ministry of Enterprise and Made in Italy, in partnership with Cineca.
Expert.ai, a leading provider of enterprise AI solutions for business value creation, announces the availability of the alpha version of the first LLM entirely “made in Italy” designed specifically for the machinery sector. The model, named EVA (Engineering & Value Automation), developed with UCIMA, the Italian Association of Manufacturers of Automatic Packaging Machinery, on behalf of the MADE Competence Center (selected by the Ministry of Enterprise and Made in Italy for Industry 4.0), is a strategic milestone for national industrial innovation. It promotes the practical adoption of AI in production processes through resources designed to simplify the digital transformation of SMEs, ensuring greater efficiency and sustainability.
Certified expertise, built by industry for industry
A wealth of complex data and information provided by more than 20 companies in the manufacturing sector forms the knowledge base used by Expert.ai to train UCIMA’s EVA model. Organizations specializing in machinery manufacturing contributed directly and systematically to the development of an AI platform designed to serve the industry and address its operational, decision making and governance needs by sharing technical documentation, best practices, processes and specialized expertise.
EVA operates on reliable, consistent data validated by the industry itself, overcoming one of the most critical limitations of generalist models: the lack of context and vertical specialization. It has also been tested in the field, with the goal of capturing not only information, but also the operational know-how that defines manufacturing, an intellectual asset built over decades of experience that often remains informal and therefore difficult to transfer or codify into AI systems.
Ready-to-use AI and technological sovereignty
The EVA LLM is already available in alpha version and ready for use in real-world applications.
Testing and optimization will continue over the next two months with the goal of delivering a fully optimized version leveraging the computing power, infrastructure and expertise in managing complex, computationally intensive systems of one of Europe’s leading supercomputing centers, Cineca. The strategic partnership with Cineca ensures scalability, reliability and technological sovereignty, key elements for next-generation industrial AI.
“The potential that artificial intelligence can bring to manufacturing is no longer a mystery. The difference lies in how the technology is designed,” said Expert.ai CEO Dario Pardi. “We started with the industrial reality and the people who work in production processes every day and we have demonstrated that it is possible to build an Italian AI that is specialized and truly useful, moving beyond the idea of technology as an end in itself to leverage know-how, increase business competitiveness and support a sustainable evolution of the manufacturing sector.”
RAG, Governance and Efficiency
To promote the use of the EVA LLM and accelerate AI adoption in real-world use cases, UCIMA is making a Retrieval-Augmented Generation (RAG) system available to its members. Developed by Expert.ai to simplify the retrieval of information, the system is based on the ability to understand and process natural language. Users can interact by asking any question—for example, to access machine manuals—and receive precise answers while maintaining traceability, accuracy and full control over the information, ensuring responses are always up to date.
