Routine: A Structural Planning Framework for LLM Agent System in Enterprise

作者：Guancheng Zeng, Xueyi Chen, Jiawang Hu, Shaohuan Qi, Yaxuan Mao, Zhantao Wang, Yifan Nie, Shuang Li, Qiuyang Feng, Pengxu Qiu, Yu-Jian Wang, Wenqiang Han, Linyan Huang, Gang Li, Jingjing Mo, Haowen Hu · 发表于：arXiv.org · 年份：2025 · DOI：10.48550/arxiv.2507.14447 · 被引用次数：8 · 研究领域：Computer Science

The deployment of agent systems in an enterprise environment is often hindered by several challenges: common models lack domain-specific process knowledge, leading to disorganized plans, missing key tools, and poor execution stability. To address this, this paper introduces Routine, a multi-step agent planning framework designed with a clear structure, explicit instructions, and seamless parameter passing to guide the agent's execution module in performing multi-step tool-calling tasks with high stability. In evaluations conducted within a real-world enterprise scenario, Routine significantly increases the execution accuracy in model tool calls, increasing the performance of GPT-4o from 41.1% to 96.3%, and Qwen3-14B from 32.6% to 83.3%. We further constructed a Routine-following training dataset and fine-tuned Qwen3-14B, resulting in an accuracy increase to 88.2% on scenario-specific evaluations, indicating improved adherence to execution plans. In addition, we employed Routine-based distillation to create a scenario-specific, multi-step tool-calling dataset. Fine-tuning on this distilled dataset raised the model's accuracy to 95.5%, approaching GPT-4o's performance. These results highlight Routine's effectiveness in distilling domain-specific tool-usage patterns and enhancing model adaptability to new scenarios. Our experimental results demonstrate that Routine provides a practical and accessible approach to building stable agent workflows, accelerating the deployment and ad...