A partner (Cappius) of RapidMiner recently gave a really cool webinar on how they use RapidMiner to analyze Call Center audio on the fly. They were interested in know when a caller's stress level changed and how they could provide better support through audio mining and analytics.
This wasn't a 100% RapidMiner implementation but rather a mix of R and Python scripts operationalized in a RapidMiner workflow. This use case is a typical application of several data science tools (i.e. Python and R) mashed together into the RapidMiner platform.