Pay for the Fault, Not the Flow: Label-Free In-Flow Multi-Agent Workflow Optimization
Abstract:Large language models (LLMs) increasingly construct multi-agent workflows that decompose a complex task and assign specialist agents from a pool. However, building such a workflow well remains challenging: how finely to divide the task, which agent to trust with each subtask, and when to create a new specialist are all critical decisions a workflow constructor needs to settle up front. Thus, whether each subtask succeeds remains unknown until the workflow runs. Yet, improving a workflow is costly. Locating a fault usually requires a reference answer, a graded outcome, or a trained assessor, and the fix is applied to the whole workflow through re-execution, re-search, or retraining. We propose InFlowOp, which prices every decision in one label-free cost that weighs how well an agent's competence meets what a subtask demands against how much that agent takes to run. Before execution, InFlowOp bidirectionally determines the granularity of task decomposition and agent assignment following from the cost rather than from a fixed template. During execution, InFlowOp corrects a fault with the cheapest move via the same cost that serves the workflow both as it is built and as it runs. Facing the workflow-level evaluation challenge, we introduce Braid, a benchmark whose tasks require multi-agent coordination beyond single-agent capability. Across various domains and backbones, InFlowOp outperforms single agent baselines by up to $+11.97\%$, achieving $+9.64\%$ with in-flow optimization. Our project page: this https URL.
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.01017 [cs.AI] |
| (or arXiv:2610.01017v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01017 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xuehang Guo [view email]
[v1]
Thu, 1 Oct 2026 04:01:23 UTC (2,650 KB)