As change agents, our role is to change how functionality is delivered. I have never seen the role of “stay the same agent” advertised. So, if change is the name of the game, how do you know whether any specific change you might champion will be good for the organization? Experiments and pilots are tools to understand whether a change will have the intended effect and to generate feedback. For many reasons (usually bad) not everyone sees the need to perform an experiment (controlled or uncontrolled) or pilot. It might seem easier to risk a full-scale implementation; however the extra risk of failure is avoidable. Why avoid as much risk as possible? Not every process change is an improvement, and in some cases there is no going back. In the end, the result might be messy.
Found picture today (Facebook)