We have a lot of data but, what answer do we expect from them? Will we always ask them the same question? How long should I wait for the recommendation taken from them? The answer to these questions raises the question of how (Edge, cloud, ..) and at what speed it should be processed, and who can program it. The simplification and location of machine learning programming tools and their future integration into our company structures will be a key element of this panel of technicians and experts.