The adoption of software development agents has surged with the prevalence of Large Language Models (LLMs), now available as IDE extensions and standalone applications. However, the justification for developing specialist agents remains ambiguous. This paper investigates this question in the context of business process automation, focusing on the transformation of Business Process Model and Notation (BPMN) diagrams into executable agentic workflows. Given that BPMN specifies explicit control-flow semantics, we concentrate on deterministic workflows where a fixed process model and inputs uniquely determine the executed path.
We introduce a specialist workflow for this task and compare it against generalist agents such as Roo and Cline. Our results indicate that the specialist solution produces agents that outperform generalist baselines by approximately 9-20 percentage points in tool-use exactness, 2-4x in penalty-adjusted latency, and 3x fewer tool-call errors, while reducing generation token cost by over 95% and eliminating repair iterations. We also find that generalist agents generate code inconsistently in both functionality and quality, thereby limiting their suitability for industrial settings where reliability and maintainability are essential.
Blogger's Review: This paper provides a comprehensive analysis of the advantages of specialist agents in specific tasks, particularly in business process automation. The clear experimental data demonstrates substantial improvements in accuracy and efficiency, offering valuable insights for future development. The limitations of generalist agents are convincingly highlighted, underscoring the importance of specialization in certain application scenarios.