This paper examines whether specialist agents justify their additional development effort for business process automation. The authors transform Business Process Model and Notation (BPMN) diagrams into executable agentic workflows and compare a specialist workflow with generalist agents such as Roo and Cline. The abstract reports that the specialist approach improves tool-use exactness by approximately 9–20 percentage points, achieves 2–4x better penalty-adjusted latency, produces about three times fewer tool-call errors, reduces generation token cost by more than 95%, and eliminates repair iterations. The setting is deterministic: a fixed process model and inputs uniquely determine the execution path.
No heat snapshots are available in the last 24 hours.