The situation
Operational workflows had grown by accretion. Each step made sense to the person performing it, and nobody held a description of the whole path, which made both automation and audit impossible to scope.
The constraint
Process work invites invention, and an in-house notation that only its author understands is worth very little two years later. Modelling had to sit on standards a third party could read, because the point of the exercise was to survive a change of personnel.
What I decided
I modelled against BPMN 2.0 and the APQC framework rather than inventing a local notation, and I traced each operational step through to its financial consequence. A workflow that cannot be reconciled against what it bills or settles is not finished, however tidy the diagram looks.
What shipped
End-to-end process models and a restructured set of workflows, expressed in standard notation with the decision points and handoffs made explicit.
What it produced
Workflows that could be automated or audited without being reverse-engineered first, and measurable gains in throughput where the bottleneck turned out to be a handoff rather than a task.