Execution systems using unstructured data include a memory, one or more input/output devices, and one or more processors coupled to the memory and the input/output devices. The one or more processors are configured to receive a working document comprising data, receive a workflow described according to a workflow structure, the workflow structure describing sub-processes of the workflow, routings between the sub-processes, actions that make up the sub-processes, and mappings between the data in the working document and the actions, identify a first sub-process to be performed, perform each of one or more first actions associated with the first sub-process, update the data in the working document based on one or more first mappings associated with the one or more first actions, and select a second sub-process to be performed based on the data in the working document and one or more first routings associated with the first sub-process.