Initiative overview
As part of the DaFNE research platform by the Austrian Federal Ministry of Agriculture, Forestry, Climate and Environmental Protection, Regions, and Water Management, there is an extensive collection of diverse information and complex data, particularly report files from various research projects that have been coordinated and published. To efficiently and systematically manage this data while ensuring it is easily retrievable, we utilise a chatbot powered by a vector database and a large language model (LLM).
The chatbot addresses the challenges associated with managing and accessing large collections of information from various research projects. It enhances information retrieval processes by moving beyond simple keyword matching, providing context-aware, relevant responses that can be easily accessed and utilised by stakeholders, including academia, administration, and the public.
Another key objective of the chatbot is to leverage advanced technology, implementing state-of-the-art AI components to facilitate meaningful interactions and deliver accurate responses to user inquiries. Furthermore, the project emphasises continuous development, committing to an ongoing effort to refine and enhance the chatbot's functionality based on user feedback and evolving needs. Networking and collaboration with similar projects will also be essential to promote development and share best practices in the future.
Initially available only to internal users, the chatbot will be made publicly accessible by the end of February 2026, thereby broadening its reach and usability. After its public launch, the focus will shift towards continuous improvement, incorporating regular updates and enhancements based on user feedback to ensure the chatbot remains relevant and effective in meeting users' needs.
Results, outcomes and impacts
To date, the chatbot’s impact has been driven primarily by the unprecedented speed at which it delivers information and its practical usability. Instead of searching for individual projects, opening their reports, and reading them, the chatbot shortens this process. This is particularly helpful when planning public relations work, knowledge transfer, and events on specific topics. Until now, this impact has only been reported through individual user experiences. Of course, it is hoped that the chatbot will be used as an everyday tool for administration in the future, but the general public and practitioners in agriculture, forestry, and water management should also benefit more from the scientific outputs.
