Reactive programming is a great way of developing scalable applications. But how can you implement complex use cases involving a bunch of nicely decoupled components (aka microservices)? How can you avoid losing sight of complex flows coupled together by peer-to-peer flows? How can you handle the state of long running interactions and deal with subsequent requirements around timeouts, error handling and consistency respectively long lived distributed transactions? How does the Saga-pattern help - and how can you implement Sagas? In this talk I will give my opinionated answer to these questions based on my real-life experience with customer projects. This involves leveraging workflow engines in a way fitting nicely into reactive architectures. I will discuss highly-scalable tools that embrace the reactive manifesto, use visual process models and can be hooked into the typical stacks. I will not only use slides but also do fun little live hacking sessions using open source components.