I was referring to people who were actually using the product. I've had a number of jobs in my life, including being in the services division of a product company and also in an internal consulting team of an extremely large non-tech company. I would typically interact directly with the people who are using what we were building. So I could see exactly what they needed and did not need (despite what they said they wanted) and manage that so they were happy with the end product.
I've also worked in situations you've described. My only interaction with the end user is through layers of middle management. I found that environment to be extremely demotivating and certainly no direct sense of satisfaction. One of the small benefits of being a Data Scientist right now is that a lot of managers don't really understand what it is I do so I get some interaction with the end-user from a requirements and solution validation standpoint. But I suspect in a few years Data Scientists will become invisible cogs (like everybody else) in the Enterprise IT machine.
I've also worked in situations you've described. My only interaction with the end user is through layers of middle management. I found that environment to be extremely demotivating and certainly no direct sense of satisfaction. One of the small benefits of being a Data Scientist right now is that a lot of managers don't really understand what it is I do so I get some interaction with the end-user from a requirements and solution validation standpoint. But I suspect in a few years Data Scientists will become invisible cogs (like everybody else) in the Enterprise IT machine.