Most process engines can route work, wait, and notify. When something has to be decided, they stop and hand it to a person. Workflow decides in place — reading the document, checking the rules and choosing the branch — and only stops when a person is genuinely needed.
A case arrives, moves through the steps on its own, and reaches a point where something has to be decided. What happens there is the whole difference: it either resolves inside the flow, or it waits for a person.
The flow reaches the point where something must be judged, leaves the system, and waits in somebody’s queue. What happens next depends on when that person opens it. The process is not slow because the work is hard: it is slow because it is not running.
The decision is a node like any other. The flow reaches it, resolves it and carries on — in the same run, with the rules written where everyone can read them. A person is called in for the cases that need judgement, not for every case.
Ten kinds of node. Everything a process does is one of these, and the palette is the same whether the flow has four steps or forty.
Where the process starts, and what data comes in with it.
What a person fills in. The same form on a desktop and on a phone.
Branches the flow according to what the data says.
Calls another system: the core, an ERP, a web service.
Assigns the task to a person or to a role.
The node that decides. It reads, classifies and validates inside the flow, without stepping out to ask.
Opens several branches in parallel and waits for them to close.
Scheduling: appointments, slots, availability.
Picking from a catalogue inside the process.
Closes the process, or hands it to the next one.
Tell us where it stops today and we will tell you what it takes to make it run — and if the answer is that Workflow is not what you need, we will say so.
Talk to us
