An integrated digital platform designed to optimise resources, reduce environmental impact, and support the food sector's green transition in line with the European Green Deal.
The food manufacturing sector is one of the most resource-intensive industries. This is due to its high energy, water and raw material consumption; the complexity of its production and logistics chains; and growing regulatory and social expectations regarding sustainability, transparency and environmental responsibility.
Against this backdrop, CLARUS was created: a Horizon Europe-funded project coordinated by Roberto Rocca and Gabriella Monteleone, researchers in the Department of Management, Economics and Industrial Engineering at the Politecnico di Milano. The project’s scientific coordinator is Professor Marco Taisch, also from the Department.
CLARUS addresses one of the central challenges of the European Green Deal by providing practical and reliable digital solutions to support the transition of the food and bio-production industries towards more sustainable, resource-efficient and environmentally friendly operating models while maintaining competitiveness and resilience in a global market.
The CLARUS project links sustainability in the food sector with artificial intelligence applications, developing an advanced digital platform. The platform integrates high communication and processing capabilities through the use of standardised communication protocols and shared data models. This enables the assessment of resource consumption and traceability throughout the food industry.
CLARUS's overarching goal is to optimise production and logistics resources in the food and time-sensitive by-product processing industries. This will allow companies to systematically measure, monitor, and enhance their environmental and operational performance in accordance with the European Green Deal's objectives and broader EU sustainability policies.
To achieve these objectives, CLARUS develops and integrates three main Tangible Expected Outcomes (TEOs): a Green Deal Index (GDI) for sustainability assessment, an IDS-compliant Data Space enabling trusted and sovereign data sharing, and an AI Toolkit providing advanced analytics and optimization capabilities
The Green Deal Index (GDI) is built upon the Green Deal Performance Assessment (GDPA) methodology and translates complex environmental, operational, and digitalization data into a single, interpretable indicator aligned with EU sustainability priorities. By combining multiple quantitative and qualitative indicators across different hierarchical levels (environment, enterprise, process, process segment and equipment), the GDI supports benchmarking, decision-making, and sustainability roadmapping for industrial stakeholders.
Moreover, together with the GDI, a Green Deal Ontology (GDO), as novel sustainability ontology for food domains has been developed in the project. The Green Deal Ontology has been developed for both industrial pilots, representing a set of representational primitives to model the environmental sustainability domain of knowledge created by the GDPA methodology.
A key result has been the design and deployment of a secure and interoperable CLARUS Data Space, compliant with International Data Spaces Association (IDSA) principles. The Data Space has evolved from a minimum viable configuration into a reinforced ecosystem including metadata discovery and transaction logging services, ensuring data sovereignty, transparency, and auditability. This infrastructure enables controlled data sharing between industrial partners, technology providers, and research organizations, while safeguarding sensitive industrial data and supporting compliance with European data governance principles.
In parallel, the CLARUS AI Toolkit has been developed and refined within an MLOps framework to address concrete industrial challenges across the food value chain. AI models have been designed, trained, and validated for predictive maintenance, energy and water consumption forecasting, anomaly detection, logistics optimization, and process optimization. Particular attention has been paid to explainability, robustness, and adaptability of the models, ensuring that AI outputs can be understood, trusted, and effectively used by industrial decision-makers. The integration of AI services with the Data Space and sustainability assessment tools enables continuous monitoring and data-driven optimization of industrial processes.
The solutions developed within the project were tested in real industrial settings. At Ardo, a company active in primary food processing, the optimisation of refrigeration systems and improved process monitoring contributed to measurable improvements in sustainability indicators used in the GDI, including a reduction in CO₂-equivalent emissions of up to 9.5% and an improvement in water efficiency of 21.6%. In the case of the Finnish company Honkajoki, which specialises in the valorisation of food by-products, the use of AI has contributed to achieve improvements in logistics and process efficiency, reducing raw material transportation time by 8.3% and improving energy efficiency.
In addition to technological development, the project focused on establishing a robust business ecosystem, showcasing achievements and disseminating the developed solutions. Market and competitor analyses were conducted, and exploitation strategies and roadmaps were defined for both individual key results and the CLARUS solution as a whole.
CLARUS makes a valuable contribution to current practices by integrating sustainability assessment, reliable data sharing and artificial intelligence-driven optimisation into a single, interoperable framework. This targeted approach addresses the specific needs of the food industry. The project demonstrates how standardised data models, open protocols and AI can support more informed decision-making and generate concrete benefits, including reduced energy and water consumption, greater resource efficiency, improved process traceability, and stronger alignment with European policies on sustainability and data governance.
The CLARUS architecture was designed from the outset to be both replicable and scalable. Thanks to its modular structure, compliance with European data sharing frameworks and the use of open standards, the solution can be adopted in similar contexts to food manufacturing and bio-production, as well as in other resource-intensive industrial sectors. In this way, the project supports the digital and green transition of European industry in the long term, contributing to greater competitiveness, job creation in digital and green technologies, and growing confidence in AI-based decision-making systems.