AI applied to medicine has undergone exponential growth for a number of years if we go by the curve of scientific publications in the field. But the main challenge facing Medical Artificial Intelligence remains the cooperation between the data scientist and the clinician and how to allow these two disciplines to co-exist and ensure that the experts work hand in hand on the design of algorithms, because AI in medicine can only be triggered by means of the human practitioner setting an example. Diagnostic assistance, personalised medicine and long-term patient monitoring... how can we design, develop and integrate AI into medical devices? How can we ensure product performance? Good practices and feedback!