| title | DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations |
|---|---|
| authors | Jiazhen Jiang, Boxi Cao, Lingyong Yan, Yaojie Lu, Hongyu Lin, Shuaiqiang Wang, Dawei Yin, Xianpei Han, Le Sun |
| date | 2026-07-22 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2607.19865 |
| upvotes | 7 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, revealing that even the most advanced frontier configurations still exhibit profound limitations when handling highly coupled, long-range tasks. Furthermore, a fine-grained analysis of existing agents' manipulation behaviors uncovers 3 key failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata. Ultimately, our work exposes the capability boundaries of agents in maintaining global document consistency, shedding light on the future design of robust, non-destructive agents for complex digital ecosystems. |
cite
@misc{jiang2026docops,
title={DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations},
author={Jiazhen Jiang and Boxi Cao and Lingyong Yan and Yaojie Lu and Hongyu Lin and Shuaiqiang Wang and Dawei Yin and Xianpei Han and Le Sun},
year={2026},
eprint={2607.19865},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2607.19865}
}