Empowering Smallholder Farmers with AgriLLM: A Low-Latency, High-Accuracy AI Assistant for Context-Relevant Farming Advice
Year: 2025
Type: article
Abstract: Smallholder farmers lack access to localized agricultural advice, while generic AI tools often provide unsafe or irrelevant recommendations. This paper introduces AgriLLM, an AI assistant that delivers accurate, context-aware farming guidance by combining large language models with region-specific agricultural databases. Tested on 500 queries from Indian farmers, AgriLLM achieves 89.6% accuracy (vs. 64.2% for standard GPT-4 [2]) and reduces incorrect advice by 81%. The system runs on budget smartphones with <2 sec response times, making it practical for rural areas. Field trials with 112 farmers showed 92% satisfaction, with particular praise for its crop-specific pesticide warnings and irrigation advice. AgriLLM demonstrates how AI can empower farmers when grounded in local knowledge, offering a scalable solution to global food security challenges. Future work will expand support for voice inputs and rare crops.
Author: Mrugesh Patel, Hemangini Patel
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: green
OpenAlex ID: W6912073150
FWCI:
Cites:
Cited by:
Related to: