We investigate the impact of social and governance pillars on the creditworthiness of Italian non-financial firms using the In-house Credit Assessment System (ICAS) of Bank of Italy. The social indicators are drawn from the National Social Insurance Agency database. To obtain the governance indicators we exploit a granular knowledge graph of Italian firms enriched through automated reasoning. We find evidence of discriminatory and predictive power of the S and G variables towards the event of default. In particular, these variables are significant for predicting the default of micro and small firms even after controlling for the financial and credit behaviour variables of the baseline statistical model, although their overall contribution to model performance is moderate. For medium—and large-sized firms, only the social variables retain statistical significance in the augmented model.