A business-license approval process needs multi-step orchestration with wait states and human approval callbacks before the license is issued. Which AWS service should the architect choose for that workflow?

Business-license approval needs multi-step orchestration with wait states and human approval callbacks before issuance. AWS Step Functions is designed for that workflow: state machines express sequential and parallel steps, can wait for timers or external callbacks (task tokens), and resume when a reviewer approves or rejects. For municipal licensing, that yields an auditable state history instead of ad-hoc scripts scattered across servers. The service natively tracks which step succeeded, failed, or is waiting on a human, which matches how city clerks and department heads actually approve licenses over hours or days rather than milliseconds.

Amazon S3 inventory reports list objects for storage analytics; they are not a workflow engine and cannot wait on human approvals. CloudFront Functions run short edge logic with strict time and capability limits and cannot manage multi-day human approvals or durable business state. AWS WAF rate-based rules throttle HTTP floods at the edge; they do not sequence license approval business processes or invoke downstream issuance systems.

Exam caveat: choose Step Functions for orchestration with waits and callbacks; EventBridge routes events but is not a long-running human-approval state machine by itself; SWF appears in older materials but Step Functions is the modern exam default. Operational check: model submit → validate → wait for approval → issue license, use a callback task token for the human step, test timeout and escalation paths, and export execution history for auditors reviewing municipal licensing controls.