Closing the Loop: Rapid prototyping for data science and AI products

It is often expensive to design and deliver products involving data science. The main reasons are that they are notoriously difficult to scope, estimate, and scale and can struggle to prove their generated impact quantitatively. We need new methods to alleviate these issues. We have drawn inspiration from existing frameworks in agile, lean, and design thinking to create a new method - “closing the loop”. It effectively limits the scope to a single, shortest possible, end-to-end value stream, attached to a measurable outcome to the first version before iterating further (i.e., adding new data sources or features) once we prove value delivery.