Damp and mould in buildings present persistent challenges for public health and building management, contributing to respiratory illnesses, degraded indoor air quality, and increased maintenance costs. A key challenge is the reliance on traditional detection approaches, such as manual inspections, occupant questionnaires, and short-term environmental measurements, which are often subjective, time-consuming, and reactive. For example, isolated temperature and relative humidity (RH) readings frequently fail to capture seasonal moisture fluctuations, while inspections typically identify problems only after visible mould growth has occurred, limiting opportunities for early intervention. In addition, existing monitoring systems rarely integrate environmental data with building-specific contextual information, such as construction details or historical diagnostic records, reducing their ability to identify the root causes of dampness. To address these challenges, this study proposes the Integrated Environmental Information System for Dampness and Mould Prevention (IEIS-DMP), a scalable and sustainable Artificial Intelligence (AI)-driven framework for proactive damp and mould risk management. The system integrates Large Language Models (LLMs) with an Agentic Retrieval-Augmented Generation (Agentic RAG) architecture, enabling autonomous planning, multimodal data retrieval, and contextual reasoning. IEIS-DMP combines high-resolution sensor data, Building Information Modelling (BIM), and unstructured diagnostic documents to continuously assess indoor environmental conditions. Through a Natural Language (NL) interface, users can obtain timely, evidence-based insights and targeted recommendations. Validation using real-world datasets demonstrates strong system performance, achieving 95.2% completeness and 94.6% accuracy. These outcomes show that IEIS-DMP supports early risk identification, informed decision-making, reduced remediation costs, and healthier indoor environments, while its modular design enables scalability and adaptation to other environmental monitoring domains. • AI system enables early detection of damp and mold in buildings. • Combines LLMs with Agentic RAG for autonomous data interpretation. • Integrates sensor, BIM, and diagnostic data for real-time monitoring. • Achieved 95.2% completeness and 94.6% accuracy in real-world tests. • Promotes healthy indoor spaces and reduces remediation costs.
Arslan et al. (Sun,) studied this question.