You can model almost any business workflow with ten disciplined questions. The objective is not to force every organization into the same process. It is to make the organization’s own choices explicit enough to review, implement, and improve.
Questions one through three: purpose, trigger, outcome
Why does this workflow exist? What exact event begins it? What observable evidence means it is complete? These questions establish boundaries. They prevent adjacent problems from quietly expanding the scope and keep the model tied to business value.
For employee onboarding, the trigger might be an accepted offer—not an informal hiring conversation—and completion might require active access, acknowledged policies, assigned supervision, and verified payroll data.
Questions four through six: actors, facts, decisions
Who participates and what authority does each participant hold? Which facts are required, where do they originate, and who owns their accuracy? Which decisions alter the path?
Avoid naming only departments. Identify the role acting at each moment and the conditions on that authority. Separate a recommendation from an approval and an approval from execution.
Questions seven and eight: rules and exceptions
What deterministic rules connect facts to actions? What happens when information is missing, a dependency fails, time expires, or someone disputes the result? Write rules in language that can become both a test and an explanation.
Exceptions should return to an accountable state. “Notify someone” is not resolution; name who owns the exception, what they can do, and how the workflow safely resumes or closes.
Questions nine and ten: evidence and ownership
What must be recorded to prove the work was done correctly? Who owns performance and approves future changes to the model? Evidence may include timestamps, source versions, rationales, approvals, delivery confirmations, and measured outcomes.
Apply the ten questions to one real case, then test a normal scenario, a boundary case, and a failure. The gaps that appear are the next modeling work.
Model conversion exercise
Use this exercise with a real workflow and the people accountable for its outcome. Record disagreements as modeling questions instead of silently choosing an answer.
Run a ninety-minute workshop with the ten questions as the agenda. Use one real case, not an imagined average. Put unresolved answers in an explicit decision log with an owner and due date. After the workshop, replay a normal case, a threshold case, and a failure against the draft. Track where participants infer different outcomes. Each disagreement becomes a missing rule, authority boundary, or fact definition. Close those gaps, publish version one, and nominate a model owner who can approve changes as operating evidence accumulates.