The Dydon system can be operated on premise or in a cloud depending on data and computing power requirements as well as regulatory, confidentiality and security requirements. Hybrid models are technically possible (e.g. rating in-house, AI engine on public data in a cloud) but may require data anonymisers. Dydon’s asset will not only be this software and models (e.g. for compliance), but also the knowledge data collected in its cloud data repository.