(A)I-COMPLI: AI Supported Automated Construction compliance inspection

Non-compliance with regulations and poor quality during on-site construction are significant international issues. For example, defects in numerous apartment buildings in Ireland are projected to incur substantial costs. Current on-site inspections are primarily visual and manual, with error rates of up to thirty percent. Inspections are often performed on a sampling basis, leaving many building elements unchecked. This has been linked to behavioral issues among operatives, contributing to poor quality.

This project aims to transform visual construction inspections by using Artificial Intelligence techniques, such as Computer Vision and Machine Learning, through camera technology. By utilizing reality capture from real and lab-based construction examples, our system is trained to identify and prompt potential non-compliance issues for further evaluation.

The solution provides a proof of concept for enhancing visual inspections with a time and location-specific digital tool. It has the potential to reduce defects and related costs by automatically flagging potential issues, thereby transforming compliance behavior. This innovation supports sector digitization and delivers positive social, economic, and environmental impacts.