Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ProAgent: From Robotic Process Automation to Agentic Process Automation

Yining Ye, Xin Cong, Shizuo Tian, Jiannan Cao, Hao Wang, Yujia Qin, Yaxi Lu, Heyang Yu, Huadong Wang, Yankai Lin, Zhiyuan Liu, Maosong Sun

TL;DR

RPA cannot handle tasks requiring human-like intelligence and dynamic decision-making. The authors propose Agentic Process Automation (APA) that leverages LLM-based agents to both construct and execute workflows, embodied by ProAgent. ProAgent uses an Agentic Workflow Description Language to separate data flow (JSON) from control flow (Python) and introduces DataAgent and ControlAgent to manage dynamic data processing and decision-making, respectively. Empirical proof-of-concept experiments on an open platform demonstrate feasible autonomous workflow construction and execution, with public code availability and discussion of tool integration, process mining, and safety considerations. The work offers a path toward offloading intelligent labor from humans while enabling flexible, agent-driven automation across domains.

Abstract

From ancient water wheels to robotic process automation (RPA), automation technology has evolved throughout history to liberate human beings from arduous tasks. Yet, RPA struggles with tasks needing human-like intelligence, especially in elaborate design of workflow construction and dynamic decision-making in workflow execution. As Large Language Models (LLMs) have emerged human-like intelligence, this paper introduces Agentic Process Automation (APA), a groundbreaking automation paradigm using LLM-based agents for advanced automation by offloading the human labor to agents associated with construction and execution. We then instantiate ProAgent, an LLM-based agent designed to craft workflows from human instructions and make intricate decisions by coordinating specialized agents. Empirical experiments are conducted to detail its construction and execution procedure of workflow, showcasing the feasibility of APA, unveiling the possibility of a new paradigm of automation driven by agents. Our code is public at https://github.com/OpenBMB/ProAgent.

ProAgent: From Robotic Process Automation to Agentic Process Automation

TL;DR

RPA cannot handle tasks requiring human-like intelligence and dynamic decision-making. The authors propose Agentic Process Automation (APA) that leverages LLM-based agents to both construct and execute workflows, embodied by ProAgent. ProAgent uses an Agentic Workflow Description Language to separate data flow (JSON) from control flow (Python) and introduces DataAgent and ControlAgent to manage dynamic data processing and decision-making, respectively. Empirical proof-of-concept experiments on an open platform demonstrate feasible autonomous workflow construction and execution, with public code availability and discussion of tool integration, process mining, and safety considerations. The work offers a path toward offloading intelligent labor from humans while enabling flexible, agent-driven automation across domains.

Abstract

From ancient water wheels to robotic process automation (RPA), automation technology has evolved throughout history to liberate human beings from arduous tasks. Yet, RPA struggles with tasks needing human-like intelligence, especially in elaborate design of workflow construction and dynamic decision-making in workflow execution. As Large Language Models (LLMs) have emerged human-like intelligence, this paper introduces Agentic Process Automation (APA), a groundbreaking automation paradigm using LLM-based agents for advanced automation by offloading the human labor to agents associated with construction and execution. We then instantiate ProAgent, an LLM-based agent designed to craft workflows from human instructions and make intricate decisions by coordinating specialized agents. Empirical experiments are conducted to detail its construction and execution procedure of workflow, showcasing the feasibility of APA, unveiling the possibility of a new paradigm of automation driven by agents. Our code is public at https://github.com/OpenBMB/ProAgent.
Paper Structure (23 sections, 2 equations, 10 figures, 1 table)

This paper contains 23 sections, 2 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. Agentic Workflow Description Language
  4. JSON Structure for Data Flow
  5. Python Code for Control Flow
  6. Agent-Integrated Workflow
  7. DataAgent
  8. ControlAgent
  9. Workflow Construction
  10. Workflow Execution
  11. Proof-of-Concept Experiment
  12. Task Construction
  13. Workflow Construction
  14. Workflow Execution
  15. Discussion
  16. ...and 8 more sections

Figures (10)

  • Figure 1: The comparison between Robotic Process Automation and Agentic Process Automation.
  • Figure 2: Illustration of Agentic Workflow Description Language.
  • Figure 3: Illustration of Agentic Workflow Description Language with DataAgent and ControlAgent.
  • Figure 4: The Illustration of the workflow construction procedure of ProAgent.
  • Figure 5: The Illustration of the workflow construction procedure of ProAgent for case analysis.
  • ...and 5 more figures