Virtual assistants (chatbots) are increasingly applied in agriculture to automate processes, improve decision-making, and provide personalized support. This paper investigates the development of a chatbot tailored to apiculture, aiming to support and educate beekeepers. Beekeeping in Greece is a major economic activity, with approximately 2.1 million hives and an annual production of 15,000-20,000 tons of honey. Yet, beekeepers often lack education and training in hive management, disease treatment and timely access to reliable information.

The development methodology followed a two-phase approach. First, a domain-specific knowledge base was compiled from certified sources, including educational and scientific literature, extension services, and European and national guidelines on beekeeping. This corpus was then used to fine-tune a general large language model to ensure accuracy in responding to apiculture-related queries. Second, the chatbot was tested with two user groups: experienced beekeepers and students in agricultural sciences. Experts evaluated the technical accuracy and reliability of responses, while novices assessed clarity, accessibility, and learning support. Key indicators included precision of answers, user satisfaction, and the tool’s ability to adapt to different levels of expertise.

Results show that the chatbot successfully addressed general management issues such as hive inspection, seasonal feeding, and honey harvesting. It also proved particularly useful in clarifying terminology, providing step-by-step guidance, and reinforcing theoretical knowledge. Novice users valued the chatbot as a self-learning tool that complemented lectures, offered immediate feedback, and allowed them to practice problem-solving in simulated scenarios. For professional beekeepers, the chatbot provided quick access to targeted information, though expert validation remained essential in complex cases of disease diagnosis. Fine-tuning with domain-specific data significantly reduced irrelevant or inaccurate responses compared to a baseline model, demonstrating the importance of focused training.

The study concludes that chatbots can be a valuable resource in apiculture by combining practical support with educational benefits. For beekeepers, they offer continuous training opportunities, easy access to certified knowledge, and the ability to refresh skills outside formal courses. For agricultural students, they provide an interactive and personalized learning environment, promoting self-directed study, digital literacy, and deeper engagement with apiculture topics. Beyond individual benefits, such tools can enhance lifelong learning in rural areas, where traditional training opportunities may be limited. Future work includes expanding the knowledge base, and embedding the chatbot in agricultural training platforms at universities and vocational schools. Overall, virtual assistants can bridge innovation and education, contributing to more sustainable and knowledge-driven apiculture.

Acknowledgement:
This publication is part of the TALLHEDA project that has received funding from the European Union's Horizon Europe research and innovation programme under grant agreement No. 101136578. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency (REA). Neither the European Union nor the granting authority can be held responsible for them.

Keywords: Chatbots, Virtual assistants, Farmer training, Education technology, Beekeeping, Agricultural technology, Natural Language Processing, Decision Support Systems.