Process Modeling and Automation
What you must know
A process model is a sequence of nodes: start/end events, user input tasks, service tasks (write to data store, integration, script), gateways (XOR for exclusive paths, AND for parallel), subprocesses, and timer events. Understand process variables (process-wide, typed) versus node inputs/outputs, and how activity class (rules, integrations, etc.) determines what a task can do.
Know process instance security: who can view, who can act on a task, and how assignment works (user, group, or expression). Understand escalations (time-based reassignment or notification) and exceptions (error handling paths, alerting). Know the difference between a synchronous subprocess (waits) and an asynchronous subprocess (does not block the parent).
Traps
- Assuming an XOR gateway evaluates all outgoing paths; it takes the first true condition.
- Forgetting that a user input task must have an assigned user or group, or it stalls.
- Confusing process variable scope with node input scope; a node cannot see a variable it was not given.
- Believing a timer event fires only once; recurring timers need configuration.
- Overlooking that process model security and task assignment are separate concerns.
Rule of thumb
Model the happy path first, then add one gateway per decision point and one exception path per service call. If a question describes a "wait until" behavior, the answer is usually a timer or a receive-message event, not a user task.
Worked example
An approval process: Start → User Input (manager) → XOR gateway. If approved, Service Task (write to data store) → End. If rejected, User Input (requester) → End. Add an escalation on the manager task after 2 days that reassigns to the manager's manager.