Periodic inspection of the integrity of welds is essential for maintaining safe and efficient operation of critical assets. Methods based on ultrasound have numerous advantages – they are fast, accessible and non-disruptive – but they meet serious obstacles when used on thick welds with a complex microstructure. The H2020 ADVISE project developed and demonstrated techniques to address these challenges in the nuclear energy context. For most of these, the knowledge of the internal material structure of the weld is essential. Building up foundations laid by ADVISE, iWeld aims to progress towards implementing the imaging and characterisation methods in-situ and broaden their application scope beyond the nuclear energy sector. The project will combine insights from a wide spectrum of industries to construct a comprehensive weld library which will be used to train an AI model predicting the internal material structure of the weld. These predictions will be used in further refined inversion and imaging algorithms to deliver a comprehensive inspection planning workflow. The ambition of iWeld is to transform the way to perform ultrasonic inspection on complex welds across several branches of industry and to deliver open tools to support their development.