Salem J. AbabnehHow I workMethod 04
From a process map to something that actually runs
The fastest way to make a bad process permanent is to automate it. The work is in deciding what should run automatically, what needs the underlying system changed, and what is better left alone.
The method
Mine the real path before modelling the intended one
The documented process and the executed process are rarely the same. The gap between them is usually where the actual problem lives.
Model in BPMN 2.0, with automation in mind
A standard notation means the model survives the person who drew it.
Redesign before automating
Remove the steps that exist only because of an old system constraint or a control that no longer applies. Automating those cements them.
Then choose the mechanism on volume, judgement and stability
Three questions decide it. How often does this run, how much judgement does a step genuinely need, and is the process stable or about to change. High volume, low judgement and stable is the only combination that should be automated outright. If a step exists only because a system cannot do something, the answer is to change the system, not to wrap a robot around the workaround. If volume is low, judgement is high, or the process is about to be replaced, leaving it manual is the correct answer and should be written down as a decision rather than left looking like an omission.
Never automate an unstable process
If the path is still changing month to month, automation freezes a moving target and you pay to rebuild it. Stabilise first, then automate.
Put an SLA on it and watch it
An automation nobody monitors fails silently, and is discovered by the customer.
Where this comes from
86 bank processes automated end to end, a service-request tracking system covering more than 300 banking processes, real-time SLA dashboards, and process reengineering to BPMN 2.0 with process mining across government entities.
The other methods
Changing how a large organization works? Let's talk.