## Inland Water Monitoring with AI Inland water monitoring is vital for public health and ecosystem protection. A new artificial intelligence system, called NAIAD, is proposed as a comprehensive solution for this task, leveraging Earth Observation data. NAIAD utilizes Large Language Models (LLMs) and external analytical tools to provide detailed information on the state of inland waters. The system is designed to be accessible to both experts and non-experts, offering an intuitive interface based on natural language queries. ## Functionality and Performance The system integrates various data sources, including weather data, Sentinel-2 satellite imagery, remote sensing indices, and established platforms such as CyFi. NAIAD uses Retrieval-Augmented Generation (RAG) techniques, LLM reasoning, external tool orchestration, and computational graph execution to process user requests and generate tailored reports. Preliminary evaluations show that NAIAD achieves over 77% correctness and over 85% relevancy on a dedicated benchmark. The results also highlight the system's adaptability and robustness to different types of queries. In particular, the Gemma 3 (27B) and Qwen 2.5 (14B) models offer a good balance between computational efficiency and reasoning capabilities. ## General Context The use of artificial intelligence in environmental monitoring is becoming increasingly widespread. Systems like NAIAD can help improve water resource management and prevent risks to public health and the environment. The ability to analyze large amounts of data from various sources and provide real-time information represents a significant advantage over traditional methods.