Havant and Pontenossa: an italian model for smart manufacturing processes

The collaboration between Havant—the successor to SB Italia—and Pontenossa SpA, a Bergamo-based leader in metal recovery and the conversion of industrial waste into secondary raw materials, stands today as one of the most significant examples of the advanced application of artificial intelligence in the Italian metallurgical industry. The model developed has leveraged the company’s information assets, the expertise of its people, and the ability of AI to support and amplify technical know-how, generating tangible benefits in terms of efficiency, reliability, and sustainability.

The project was launched with the goal of optimizing the production process by analyzing and correlating data from sensors distributed throughout the plant with data from chemical analyses of the materials used in the process, thereby translating the plant’s operational complexity into tools capable of supporting operators in their day-to-day decision-making. The importance of the information assets has been a defining feature from the very beginning: sixteen years of historical data—collected, cleaned, and validated—have enabled the construction of predictive models grounded both in the depth of the data and in scientific knowledge, as well as the modeling of physicochemical, thermodynamic, kinetic, and heat transfer processes. As Claudio Cerioli, CEO of Pontenossa, emphasizes, “this is not a simple technological upgrade, but a strategic investment to strengthen internal expertise: AI amplifies our ability to interpret the production process and anticipate its trends thanks to the combination with the experience of our staff.”

Havant played a decisive role in establishing a structured approach focused not only on developing predictive models but also on building a true digital ecosystem. The company brought to Pontenossa its ability to integrate data interpretation, AI tools, document management solutions, and advanced analytics platforms. The work was conducted in close collaboration with technicians, operators, and plant managers, with the goal of gaining a detailed understanding of the production process and translating operational know-how into models capable of generating immediate value. Thanks to this approach, Pontenossa now has a coherent and scalable technological architecture in which certified data, predictive algorithms, and operational dashboards interact natively and provide reliable information for every level of the organization.

“This project demonstrates that artificial intelligence is not an accessory, but a strategic tool for building a more resilient, sustainable, and competitive industry,” says Pietro Lanza, General Manager and CEO of Havant. “Our strength lies in transforming operational complexity into tools that speak the language of operators and generate immediate value. When data becomes reliable and accessible, AI can truly amplify people’s expertise and support decisions that improve product quality, energy efficiency, and environmental impact.”

Pontenossa’s journey began with Fenice I, an initial predictive model designed to support the operation of the Waelz furnace, which extracts zinc in the form of oxide from a byproduct of steel production. By simultaneously analyzing thousands of plant and laboratory parameters, the model provides operators with precise and timely recommendations, helping to reduce downtime, improve anomaly management, and stabilize the entire process. The project was quickly adopted as both an operational tool and a training aid, thanks to the oversight of Production Manager Luca Re, who ensured full alignment with the department’s needs.

Based on the results obtained, the project was expanded to Fenice II, which focuses on extraction yield and reagent consumption. This evolution has improved the consistency of results, increased volumes, and reduced variability, confirming AI’s ability to generate value when it operates on solid data and through continuous dialogue with process experts.

At the same time, Pontenossa launched a process to digitize administrative and procurement processes using Docsweb, Havant’s Smart Enterprise Content & Workflow Management platform. The introduction of a digital, traceable, and compliant management system has improved the efficiency of internal workflows and made the relationship between operational areas and support functions more integrated.

Furthermore, the project also led to the development of the corporate DataHub, now the reference source for all data generated and collected at the Pontenossa site. The DataHub certifies and distributes this data across various internal domains, enables the development of new AI applications, and forms the foundation of a robust, scalable, and future-oriented data governance framework. Thanks to integration with Business Intelligence tools such as Qlik, information is now easily and immediately accessible at all levels of the company, supporting strategic analysis and daily operational oversight.

The project has also yielded tangible results in terms of sustainability. By 2025, Pontenossa had increased its zinc extraction yield by 15%, reduced reagent consumption by 9%, and decreased the volume of production residues. The company is now expanding its roadmap to include the creation of advanced energy KPIs, mass and energy balances integrated into the DataHub, and the development of optimization models aimed at defining the most efficient recipes and energy mix. “Innovation also means being more responsible toward the environment,” says Cerioli. “AI has allowed us to reduce consumption and emissions and improve the quality of the residue, confirming the link between operational efficiency and sustainability.”

The next step will be the completion of the Fenice project as a unified tool for operators, capable of integrating predictive models, certified data, and operational dashboards within a single interface. The roadmap calls for, over time, expanding the scope of information to include energy consumption, developing predictive maintenance models, implementing proactive anomaly management, and establishing an ongoing training program dedicated to both traditional and generative AI. The transformation, in fact, has been not only technological but above all cultural. “Understanding that artificial intelligence does not replace but rather enhances our capabilities was a decisive step,” concludes Cerioli. “This is why investment in training is essential, so that every operator knows how to interpret the data and use it as a lever for improvement.”

The collaboration between Havant and Pontenossa demonstrates how a robust data culture, combined with in-house expertise and scientific rigor, can generate structural benefits and strengthen the role of Italian industry in a competitive and sustainable context.