Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.08158 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909988647075840 |
|---|---|
| author | Zhou, Yuqing Wang, Zhuoer Yuan, Jie Wang, Hong Koelle, Samson Zhu, Ziwei Niu, Wei |
| author_facet | Zhou, Yuqing Wang, Zhuoer Yuan, Jie Wang, Hong Koelle, Samson Zhu, Ziwei Niu, Wei |
| contents | Large language model (LLM)-based agents are widely deployed in user-facing services but remain error-prone in new tasks, tend to repeat the same failure patterns, and show substantial run-to-run variability. Fixing failures via environment-specific training or manual patching is costly and hard to scale. To enable self-evolving agents in user-facing service environments, we propose WISE-Flow, a workflow-centric framework that converts historical service interactions into reusable procedural experience by inducing workflows with prerequisite-augmented action blocks. At deployment, WISE-Flow aligns the agent's execution trajectory to retrieved workflows and performs prerequisite-aware feasibility reasoning to achieve state-grounded next actions. Experiments on ToolSandbox and $τ^2$-bench show consistent improvement across base models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_08158 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | WISE-Flow: Workflow-Induced Structured Experience for Self-Evolving Conversational Service Agents Zhou, Yuqing Wang, Zhuoer Yuan, Jie Wang, Hong Koelle, Samson Zhu, Ziwei Niu, Wei Computation and Language Large language model (LLM)-based agents are widely deployed in user-facing services but remain error-prone in new tasks, tend to repeat the same failure patterns, and show substantial run-to-run variability. Fixing failures via environment-specific training or manual patching is costly and hard to scale. To enable self-evolving agents in user-facing service environments, we propose WISE-Flow, a workflow-centric framework that converts historical service interactions into reusable procedural experience by inducing workflows with prerequisite-augmented action blocks. At deployment, WISE-Flow aligns the agent's execution trajectory to retrieved workflows and performs prerequisite-aware feasibility reasoning to achieve state-grounded next actions. Experiments on ToolSandbox and $τ^2$-bench show consistent improvement across base models. |
| title | WISE-Flow: Workflow-Induced Structured Experience for Self-Evolving Conversational Service Agents |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2601.08158 |