AI4ESOPP: AI for More Sustainable Production
A project developed by Mashfrog that integrates industrial data and AI to improve production planning and resource utilization.
Planning industrial production means balancing orders to fulfill, deadlines to meet, and available resources every day. Factoring energy consumption and other external variables into these decisions opens up new opportunities: choosing when to carry out a production operation can help make it more efficient and sustainable.
This is the context in which AI4ESOPP (Artificial Intelligence for Environment and Sustainability of Production Plants) was developed. This Research and Development project by Mashfrog Group was carried out under the “Entrepreneurial Discovery” measure of the Sustainable Growth Fund, Action 1.1.4 “Collaborative Research” of the PN RIC 2021–2027 program, promoted by the Ministry of Enterprises and Made in Italy (MIMIT). The goal is to apply artificial intelligence to industrial plant management, integrating data, models, and optimization tools to support production planning and environmental sustainability.
Five Functional Components, One Goal
The AI4ESOPP platform can be viewed as a combination of five functional components:
-
Data collection and management: a component designed to collect, aggregate, and make accessible all available data, both internal and external, relating to the production plant;
-
Monitoring and visualization: a dashboard that enables different types of analysis depending on the type of plant;
-
Modeling, simulation, and optimization;
-
Environmental sustainability and certifications: an advanced reporting component that supports the company in systematically, objectively, and periodically assessing its environmental impact for ESG (Environmental, Social, and Governance) assessments, monitoring environmental performance KPIs, and supporting processes related to LEED, ISO 50001, ISO 14001, and EMAS;
-
LLM-based virtual assistants to support users in interacting with the platform and accessing data and information through natural language.
AI makes it possible to connect information that influences the same decision: what to produce, by when, and using which resources. The expected benefit is more informed planning, in which production and environmental goals are evaluated together.
Adding an AI Layer to SAP Planning
The involvement of D’Amico - an Italian company specializing in the production of jarred and canned foods - as an industrial partner enabled Mashfrog to work with real industrial production processes. The production plan available in SAP served as the basis for evaluating alternative production sequences and schedules, taking additional efficiency and sustainability criteria into account while meeting order deadlines.
The AI4ESOPP platform, which uses the SAP production plan as its input and adds an artificial intelligence layer, was able to support production plan optimization. For example, at a plant equipped with solar panels, integrating forecasts for a sunny day into the system can provide useful information for optimizing production. In line with the production plan, energy-intensive operations can then be scheduled for times when more solar energy is expected to be available.
The production plan available in SAP served as the basis for evaluating alternative production sequences and schedules, taking additional efficiency and sustainability criteria into account while meeting order deadlines. As part of this effort, AI4ESOPP also adopted Databricks for data management and processing to support its artificial intelligence models.
Within the project, Databricks serves as the data layer between industrial sources, including SAP, and the analysis and optimization models developed by AI4ESOPP. This choice was further validated by developments within the SAP ecosystem itself: in February 2025, SAP and Databricks announced a strategic partnership to integrate data management and AI capabilities within SAP Business Data Cloud.
From Data to Everyday Decisions
To support this approach, AI4ESOPP can integrate information from enterprise systems, machinery, sensors, and external sources. Using this foundation, the platform develops analyses and predictive models to identify inefficiencies and opportunities for improvement, including in maintenance management.
The project also includes environmental performance monitoring and support for certification processes. Sustainability thus becomes part of industrial process analysis through indicators that help track changes over time.
A virtual assistant powered by a language model makes it easier to interact with the platform and access information through natural language. The goal is to help users understand the data and use it in their assessments.
Mashfrog’s Expertise Across Industry and Research
The Mashfrog R&D team brought together expertise in AI, data management, industrial systems integration, and SAP technology. This work was complemented by the scientific contribution of Roma Tre University to the research activities - with a particular focus on data management, AI-based modeling and optimization - strengthening the exchange between industrial expertise and scientific methodologies.
“With AI4ESOPP, we worked to make sustainability a tangible criterion in industrial planning,” explained Antonio Grillo, Research & Product Innovation Officer at Mashfrog Group. “We were able to demonstrate the practical value of artificial intelligence by connecting it to real production data, systems, and constraints.”