Major disasters (e.g., Ponte Morandi, 2018) highlighted the critical condition of many European civil infrastructures due to inadequate maintenance, unforeseen operating conditions, and extended usage. Regular inspections of all types of infrastructures are essential to assess their health, guide maintenance and repair, and extend service life. SARAH adopts a human-centred approach to develop a safe-by-design integrated digital solution that supports maintenance and repair decision and planning, order, intervention and reporting. Leveraging metaverse technologies and egocentric augmented reality solutions, the system empowers infrastructure supervisors and intervention workers by providing real-time assistance, fostering skill development, and improving their Occupational Safety and Health. The solution, rooted in Human Factors: a/ identifies issues in civil infrastructures in a rapid, cost-effective and safe manner thanks to innovative sensors deployed on-site (corrosion, soil nail and anchor, structural crack) possibly via an unmanned aerial system (load up to 1kg); b/ fastly and accurately assesses the needs for repair (98 % of issues properly ranked) using a Digital Twin of the infrastructure able to run in real-time (reduced order, automatic continuous calibration) and proposes prioritised timely interventions, considering risk assessment using a Decision Support Engine (80% ad hoc correct advice for inspections); c/ launches extra inspections (either with autonomous aerial system or human), and provides remote support to the intervention workers (80% satisfaction, evaluated via interviews); d/ delivers certification report generated using generative AI, after intervention and repair (at least 2 types of reports generated). With this approach, SARAH offers a game changer for infrastructure inspection, maintenance and repair.