Great question! I worked with two approaches:
1. Sagas - a centralised place to handle bussines flows
2. Event system
Start with 1. and eventually (pun intended) move to 2. Both ways allow parallelized interactions with external systems, deferred decisions, etc. whatever the bussines will require.
Sorry I wasn't clear, I'm talking about imperative code that coordinates multiple external actions. In a world where it's modeled as jobs in a distributed system, I agree that each job can be nicely functional (and I love this pattern).
I can return a tree of lambdas but then I have to resolve them against something and that's just replacing mocks with lambdas, really. Not sure it's any better in practice.