跳到正文
arXiv cs.AI· Luyao Wang, Xijun Zhu, Zhaohui Hao, Xiwen Zhang, Xingyu Wen, Weishan Li, Yuyang Liao, Yanqin Gao, Yongsheng Du, Lingjun Huang, Shaowei Zhao, Gang Li, Qifan Yu, Baoli Wang·· 4 小时前精选AI 评分65

Co-Sight Harness:用外部化控制资产和确定性执行提升长程业务智能体可靠性

Co-Sight Harness: Externalized Control Assets and Deterministic Enforcement for Long-Horizon Business Agents

AI 导读

arXiv 论文(2611.17696)提出 Co-Sight Harness 方法,在不更换语言模型的前提下,把工作流定义、状态和领域知识外部化为可版本化的控制资产,用状态机和机器可校验的验收门在执行中强制约束,并由独立审计器评估轨迹,资产更新需回归检查与人工批准。

推荐理由

论文提出不改模型、靠外部化控制资产与确定性执行提升长程业务智能体可靠性,并给出三个超200步真实工作流的量化结果。

正文 · 原文

Authors:Luyao Wang, Xijun Zhu, Zhaohui Hao, Xiwen Zhang, Xingyu Wen, Weishan Li, Yuyang Liao, Yanqin Gao, Yongsheng Du, Lingjun Huang, Shaowei Zhao, Gang Li, Qifan Yu, Baoli Wang

View PDF HTML (experimental)

Abstract:Long-horizon business workflows fail in two ways: errors accumulate as a chain of activities lengthens, and the knowledge an activity needs is written in a different vocabulary from the user's request. We study whether such workflows can be made reliable without changing the language model. Our method keeps the backbone fixed and moves the workflow definition, state and domain knowledge into versioned control assets outside the model. A state machine and machine-checkable acceptance gates enforce these assets during execution; an independent auditor grades the resulting trace. The assets are updated only after a regression check and human approval .We evaluate the method on three deployed workflows exceeding 200 steps. On Domain A, a 120-case end-to-end evaluation set shows unattended completion rising from 50% to 100%, retrieval recall from 56% to 91.67%, and end-to-end latency dropping from 440 to 159 minutes, with scenario-analysis recall rising from 45% to 93%. Two other domains were taken over by teams that did not write the assets, one in four days and one in seven days. A controlled ablation separates the two failure sources. The code is available at this https URL.
Subjects: Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.17696 [cs.AI]
  (or arXiv:2611.17696v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.17696

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Xiwen Zhang [view email]
[v1] Fri, 13 Nov 2026 03:21:33 UTC (2,322 KB)

来源:arXiv cs.AI · arxiv.org