Actually, I guess I should be clearer. The management is actually calling this phase "Parallel Use" not "Parallel Testing".
So I would assume that all QA testing would be completed prior to this Parallel Use phase. I guess what happened during our last Parallel Use phase was the users noticed some strange behavior and felt that they then had to become QA testers. Properly recording the behavior and reporting it.
When conducting Parallel Testing, is it best to have the actual users of the new system conduct the testing? Or should QA run the testing? The users say that they aren't QA engineers, but there is no documentation on how the system will actually be used. So QA says how can we simulate actual use? There is documentation on specific functions/features of the system, but nothing that defines how the different features will be used concurrently and which specific users will be involved.