Using LLMs to bridge the gap between consumers and Blockchain on a Agro-food traceability platform: an architectural proposal

This paper explores architectural options for enhancing farm-to-fork traceability in the agro-food industry through the integration of blockchain, Web technologies, and Large Language Models (LLMs). Our primary objective is to propose an innovative architectural approach that addresses the challenges of implementing comprehensive traceability systems while improving consumer engagement. We examine various technological components, including blockchain, LLMs, and Web technologies. By leveraging LLMs enhanced with Retrieval-Augmented Generation (RAG) and Reasoning and Acting (ReAct) frameworks, we seek to simplify user interactions with the traceability system, making complex supply chain data more accessible and understandable to consumers. This approach to farm-to-fork traceability has the potential to significantly enhance transparency, food safety, and consumer trust in the agro-food sector.