Map what actually happens
Begin with a real recent case, not the official procedure. Capture the trigger, every handoff, decision, waiting period, system, document, exception, and final outcome.
Measure the hidden work
For each step, record active time, waiting time, frequency, error rate, and who owns the next move. Waiting and follow-up often cost more than the visible task.
- Trigger and desired outcome
- People and systems involved
- Inputs, decisions, and outputs
- Active time versus waiting time
- Exceptions, rework, and escalation
Design the future state
Remove unnecessary steps before automating. Then assign one source of truth, explicit ownership, automatic routing, visible exceptions, and the few metrics that prove whether the new process is better.