AgroGenio is an AI-powered agricultural knowledge assistant developed for El Salvador under the MERIAN project and built on GENIE.AI, ITU’s open-source AI framework.
Its main purpose is to enable agricultural extension workers and advisors to quickly access practical guidance drawn from previously selected and curated technical documentation.
It minimises hallucinations combining Retrieval Augmented Generation (RAG) and Large Language model (LLM).
RAG allows AgroGenio to answer agricultural questions using curated technical knowledge relevant to El Salvador. RAG finds the relevant trusted knowledge; the LLM understands the question and turns that knowledge into a useful answer. Its primary users are
agricultural extension workers and advisors supporting farmers in the field .
Agricultural publications from CENTA and ENA are identified and curated, together with FAO publications. This content is digitally prepared and used to feed AgroGenio so that relevant information can be retrieved when a extensionist needs to address a specific agricultural question.
AgroGenio is not intended to replace institutional technical knowledge. Rather, it seeks to make that knowledge much more accessible at the point where technical assistance is provided.
AgroGenio currently has web and mobile access.
AgroGenio inherits the flexibility of its originating framework: Genie.ai:
- The LLM can be anyone you like, either open source or proprietary.
- It can evolve towards accepting different types of sources: text, images…
- It can be deployed in premises, in the cloud or hybrid, quickly.
- It can escalate rapidly to adjust to demand.
- AgroGenio is an agricultural adviser application mainly because of its corpus of knowledge.