Livestocker AI development

Detailed project description

The aim of the LIVESTOCKER - AI project is to develop a state-of-the-art decision support system based on artificial intelligence (AI) for livestock farming in Hungary and internationally. The project is implemented in a consortium by Mangal Ilona Kft. and Livestocker Solutions Kft.: the former participates as professional reference and data owner, the latter as software developer and system provider. The development builds on the existing LIVESTOCKER livestock management system, which the project takes to a new level by extending it with an intelligent assistant based on language models, predictive analytics components and autonomous decision support functions.

The main components of the system are: a web and mobile interface extended with an interactive, chat-based AI assistant, with data visualization and reporting capabilities; large language model (LLM) integration for processing structured queries and natural language interactions; a backend and data integration layer based on Function Calling and Retrieval Augmented Generation (RAG); and a Big Data and predictive analytics module that uses machine learning to process sensor data (e.g. RFID gates), behavioral patterns and economic indicators. The AI layer does not just answer questions: it automatically analyzes data, warns of anomalies, forecasts the growth curves of the animals and optimizes farm operations.

A key scientific novelty of the development is a multi-level architecture and an autonomous verification subsystem designed to filter out hallucinations (the phenomenon of AI giving incorrect or fictitious answers), with the goal of reducing the hallucination rate from 4% to below 1%; in addition, a standalone AI system ensures data quality and data consistency by filtering out erroneous data and applying alert thresholds.

The project offers a market-ready solution, under a license-based business model, for large-scale livestock farms, family farms, integrators, feed manufacturers, meat industry players, as well as agricultural universities and research institutes. The development is carried out with the consortium partners' own capacities: Livestocker Solutions Kft. provides the development, database and integration expertise, while Mangal Ilona Kft. contributes farm and technological know-how as well as testing at its reference farm. The long-term goal is to enter the international market and to adapt the technology to further animal species.

The project was implemented with the support of the European Union, co-financed by the Hungarian State.