Multi-User MIMO (MU-MIMO) systems have gained numerous attention from researchers in the past decade due to its substantial gains in the system throughput. Most of the initial research assumes the knowledge of perfect channel state information at the transmitter (CSIT). However, this is considered impractical and research over the past few years has been focused on receivers feeding back limited information to the transmitter. In this work, we consider a transparent MU-MIMO system model with limited feedback. We propose a general framework for dynamic MU-MIMO scheduling with the capability to switch between Single-User MIMO mode and (Multi-Rank) Multi-User MIMO mode without users feeding back additional multi-user information. Specifically, this is done by each user carefully estimating its CQI under the hypothesis of Multi-Rank MultiUser MIMO transmission and taking advantage of the codebook structure. We consider the sum rate of the system assuming each user uses ML and LMMSE receiver and show that our proposed scheduler, with significantly reduced feedback load, outperforms the best-companion user pairing.