Air pollution is a widespread problem due to its impact on both humans and the environment. Providing decision makers with AI-based solutions requires to monitor the ambient air quality (AQ) accurately and timely, as AI models highly depend on the underlying data used to justify the predictions. Unfortunately, in the urban context, the hyper-locality of AQ, varying from street to street, makes this difficult to monitor using high-end sensors, as the cost of the amount of sensors needed for such local measurements is too high. In addition, development of pollution dispersion models is challenging. This pilot will: i) Explore the use of machine learning (ML) for air quality prediction in the city of Trondheim, Norway, using air quality data captured by IoT devices; ii) Improve data quality and services by combining pollution data with other information (such as mobility patterns, weather forecasts, environmental data); iii) Demonstrate exploitation of the AI4EU platform. The pilot will address research challenges tied to: i) Development of reliable forecasting models based on various sensor setups and data sources; ii) Detection of anomalies originating in sensor data (missing data, noisy data, drifting sensors, etc.); iii) Visualization and explanation of models and results to decision makers (decision support for citizens, city planners, etc.). The pilot will address these challenges by building AI components with links to activities in WP7, in particular the tasks on Physical and Explainable AI. The anticipated impacts of the pilot will be a better understanding of AI capabilities for IoT, but also proof of concepts for more precise, real-time measurements of air pollution and data driven decision tools. A wider motivation for the pilot and more detailed information is available in a separate white paper.