If you want to deploy machine learning – and almost everybody does – then you need an environment that facilitates that. IBM refers to this by saying that you can’t have artificial intelligence without an information architecture (“AI requires IA”). And the problem with building an information architecture is that it involves many moving parts, many software requirements and many personas. To make this work requires that companies adopt AnalyticOps as a principle, and this requires not just a broad range of base functionality but collaborative support across all of the personas involved. Even though ICP for Data is still developing you can see that this is the direction in which the product is headed. It would be infinitely harder to achieve with a set of disparate products from multiple vendors.