Making complex operational knowledge instantly accessible is one of the most promising applications of artificial intelligence in public infrastructure management. In this presentation, Seifeddine Bettaieb showcased a collaboration between the Luxembourg Institute of Science and Technology (LIST) and the Intelligent Transportation Systems (ITS) department of Dublin City Council, demonstrating how AI can support professionals in accessing critical information quickly, securely and accurately.
Large organisations accumulate knowledge in many different formats, including documents, manuals, videos, screenshots and other digital resources. As this knowledge base continues to grow, finding the right information at the right time becomes increasingly challenging, making it more difficult to preserve and transfer valuable expertise.
The presented solution is a secure AI assistant that enables staff to ask questions in plain language and receive precise, source-based answers drawn instantly from the organisation's complete knowledge base, regardless of format. Built on a Retrieval-Augmented Generation (RAG) architecture, the assistant grounds every response in verified internal content rather than generating unsupported answers—an essential requirement in operational environments where accuracy and reliability are critical.
Security formed a key design principle throughout the project. The system operates entirely on-premise without requiring an internet connection and incorporates dedicated safeguards to ensure interactions remain secure, well-defined and fully auditable. These characteristics make it particularly well suited for deployment within public sector organisations.
A distinguishing feature of the project is that it was developed and validated using real operational content in all its complexity, rather than a simplified research dataset. The presentation demonstrated that compact, efficient AI models—when combined with the right architecture and appropriate safeguards—can deliver significant practical value without the costs or privacy concerns often associated with large cloud-based AI systems.
The session provided an inspiring example of how responsible and practical AI adoption can help public organisations improve knowledge sharing, support operational decision-making and make critical expertise readily available to the people who need it most.