WorkTeam: Constructing Workflows from Natural Language with Multi-Agents

2025-03-28

AI Summary

To address the high domain expertise barrier and degraded performance of monolithic large language models (LLMs) on complex natural language-to-workflow (NL2Workflow) translation tasks, this paper proposes a multi-agent collaborative framework—comprising a Supervisor, Orchestrator, and Filler—that decouples task understanding, workflow orchestration, and parameter instantiation, thereby overcoming LLM limitations in domain specialization and dynamic task adaptation. We introduce HW-NL2Workflow, the first real-world business-driven benchmark (3,695 samples), along with domain-adapted workflow syntax constraints and a progressive parameter-filling mechanism. Experimental results demonstrate significant improvements in workflow generation success rate on our benchmark, achieving, for the first time, high-accuracy, high-reliability end-to-end NL2Workflow deployment in enterprise settings.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Teamwork

Application Category

• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Workflows play a crucial role in enhancing enterprise efficiency by orchestrating complex processes with multiple tools or components. However, hand-crafted workflow construction requires expert knowledge, presenting significant technical barriers. Recent advancements in Large Language Models (LLMs) have improved the generation of workflows from natural language instructions (aka NL2Workflow), yet existing single LLM agent-based methods face performance degradation on complex tasks due to the need for specialized knowledge and the strain of task-switching. To tackle these challenges, we propose WorkTeam, a multi-agent NL2Workflow framework comprising a supervisor, orchestrator, and filler agent, each with distinct roles that collaboratively enhance the conversion process. As there are currently no publicly available NL2Workflow benchmarks, we also introduce the HW-NL2Workflow dataset, which includes 3,695 real-world business samples for training and evaluation. Experimental results show that our approach significantly increases the success rate of workflow construction, providing a novel and effective solution for enterprise NL2Workflow services.

Problem

• Automating workflow creation from natural language instructions
• Overcoming single-agent limitations in complex NL2Workflow tasks
• Addressing lack of specialized datasets for workflow generation

Innovation

• Multi-agent framework for NL2Workflow conversion
• Introduces HW-NL2Workflow dataset for evaluation
• Specialized agents enhance complex task performance