PSI - Issue 78

Carpanese Pietro et al. / Procedia Structural Integrity 78 (2026) 536–543

537

1. Introduction Italy ranks among the most vulnerable countries worldwide to catastrophic natural events, placing fourth in terms of average expected losses according to UNISDR (2017). Floods and earthquakes account for over 80% of the economic damages associated with these hazards, with earthquake- related costs exceeding €180 billion to the state over the past fifty years (Italian Civil Protection Department 2018). Despite this high exposure, insurance coverage remains limited to a very low percentage of incurred damages, significantly below the European average, highlighting the urgent need to strengthen effective risk management strategies. In this context, insurtech platforms play an increasingly crucial role in assessing risks and supporting the design of insurance solutions. While well-established platforms such as RMS, ImageCat, Verisk, and Oasis (https://www.moodys.com/, https://www.imagecatinc.com/, https://www.verisk.com/, https://oasislmf.org/) offer catastrophe risk modeling services worldwide, their models are often calibrated for specific regions — primarily the USA, UK, Australia, and New Zealand — limiting their applicability to the Italian context. Additionally, most of these tools operate as closed “black box” systems, offering limited transparency and interaction for users. Lastly, they often do not offer the level of accessibility, simplicity, and adaptability needed by the insurance sector. The PERIL project addresses these challenges by developing an open, interoperable, and scalable digital platform tailored to the Italian territory. It provides a multi-risk framework (NatCat model) for estimating expected losses of residential and industrial buildings from seismic, flood and other climate-related hazards, using and synergizing recent scientific advances in the Italian context (Italian Civil Protection Department 2018, Dolce et al. 2021, Perazzini et al. 2024). In addition, PERIL integrates econometric and financial models to estimate direct and indirect losses (Di Ludovico et al. 2022). An additional innovative aspect of the PERIL platform is its use of artificial intelligence (AI) technologies, such as Computer Vision, which makes it possible to remotely gather detailed building information and perform more precise risk assessments. Unlike existing tools, PERIL is designed to support insurers in generating specific financial products and insurance contracts, thus addressing current limitations in the insurtech landscape (Xu and Zweifel 2020) and strengthening resilience to natural risks in Italy. 2. Input selection and input enrichment In order to perform risk analysis in PERIL, the first step is to create a project and add buildings to it in one of the following four ways: manually clicking on the buildings of interest on the PERIL map (relying on OpenStreetMap, OSM, https://www.openstreetmap.org/); selecting all buildings in a geographical entity (town, city, province, region, etc.), or in an user-drawn polygon; uploading a file with a list of building coordinates or addresses. Fig. 1 shows the logo of the project on the left, and a screenshot of the platform input page.

Fig. 1. PERIL platform logo (on the left); screenshot of the platform inputs page (on the right).