Carbon Footprint of sourced materials calculated using AI-Driven product matching

Our client, a Dutch conglomerate of feed factories, needed to accurately assess the carbon footprint of their products. The lack of a streamlined process to combine research, procurement, and production data hindered the factories’ ability to optimize their feed recipes for minimum ecological impact. This case study highlights how Versatyle utilized AI-driven solutions to address these challenges and improve sustainability.

Challenges:

1. Accurately assessing the carbon footprint of feed products.
2. Integrating research, procurement, and production data to optimize feed recipes.
3. Enhancing sustainability efforts within the feed factories.

Solutions:

Versatyle leveraged its data expertise and AI-driven solutions to address the client’s needs effectively:

1. Integration: The Enterprise Resource Planning (ERP) system of the feed factories was integrated with the Versatyle Data Platform™ by mapping the various needed field to those matching in our Canonical Data Model.
2. Semantic Matching: A large language model was employed to semantically match raw materials from the ERP system with available research data on carbon footprints from open databases.
3. Real-time Insights: The AI model calculated the carbon footprint of each ingredient, providing real-time insights into their ecological impact. This allowed for automatic substitution with more eco-friendly alternatives during production.
4. Visual Dashboards: Versatyle established visual dashboards to track ongoing improvements. These dashboards enabled procurement to make informed decisions about sourcing materials with lower carbon emissions, further reducing ecological impact.
5. Transparency: The solution provided visual evidence of the commitment to sustainability, which was shared with customers and supply chain partners, showcasing the progress made over time.

Results: The feed factories’ ERP system is now continuously enriched with accurate carbon footprint data for every sourced material. This data-driven approach has enabled the factories to optimize their recipes, leading to a reduction in their global ecological impact.

Applicability: This technology has a wide range of potentially disruptive use cases. It facilitates large-scale matching, categorization, and deduplication of products and materials based on their name and description. Its adaptability, relying on language rather than industry-specific knowledge, makes it suitable for various business contexts.