Responsible AI - Automated testing & governance for the insurance industry Qorus-NTT DATA Innovation in Insurance Awards 2025- Nominated

Generali, the global insurance leader, is rapidly unlocking business value with generative AI, as the availability and capabilities of the technology increase. But enterprises can suffer financial, operational, or reputational damage when AI systems discriminate, provide wrong information, or seem untrustworthy. To scale AI safely and to comply with new regulations such as the EU AI Act, companies must invest in quality assurance and documentation. Generali and its Swiss innovation centre HITS have run a proof of concept with ETH spin-off Calvin Risk to automate the testing and governance of AI systems. We achieved an effort reduction of 80% compared to manual, excel-based testing and human interpretation of results of a GPT-powered customer-facing chatbot. Additionally, excel templates for project implementation and risk profiling were digitalized, enabling centralization, collaboration, and analytics. The departments involved were IT, Product, Innovation, and a 3rd party bot supplier. Competition is with either (a) manual, internal methods or (b) external tools that support either testing or governance efforts, as Calvin Risk uniquely combines testing and governance automation.