# AgentGrad: Intervention-guided Prompt Optimization for Multi-Agent Systems

DevFeed: [AgentGrad: Intervention-guided Prompt Optimization for Multi-Agent Systems](<https://devfeed.tech/articles/agentgrad-intervention-guided-prompt-optimization-for-multi-agent-systems-124475.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.08572>)

Author: Jaewon Chu, Jinwoo Seo, Jaewon Cho, Jeehye Na, Yunyang Xiong, Youngdae Kim, Hyunwoo J. Kim

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [prompt](<https://devfeed.tech/topics/prompt.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>)

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [multi-agent-systems](<https://devfeed.tech/tags/multi-agent-systems.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [performance](<https://devfeed.tech/tags/performance.md>), [prompt-optimization](<https://devfeed.tech/tags/prompt-optimization.md>), [semantic](<https://devfeed.tech/tags/semantic.md>)

## AI overview

AgentGrad improves multi-agent prompt optimization by targeting the agent responsible for each failure and grouping feedback by meaning. It injects a hint into one agent at a time; when that agent's modified output resolves the failure, the output supervises extraction of a fine-grained textual gradient. It then clusters semantically similar gradients and abstracts each cluster into a shared corrective pattern, avoiding aggregation of unrelated failures. The abstract reports state-of-the-art results on five multi-agent benchmarks, with 2.5x less wall-clock optimization time and 21.8% lower average optimization cost than the next-best baselines.

## Source excerpt

arXiv:2609.08572v3 Announce Type: replace Abstract: Large language model (LLM)-based multi-agent systems (MAS) achieve strong performance by employing specialized multiple agents, yet their performance depends on the prompt design of each agent. For MAS prompt optimization, textual gradient methods that guide prompt updates using natural-language feedback have emerged as a leading paradigm. In this paper, we identify limitations in two stages of existing textual gradient approaches: gradient extraction and gradient aggregation. In gradient extraction, previous works select a prompt to update without verifying whether modifying it resolves the failure, and derive gradients without agent-level supervision over the corresponding agent's intermediate output. In gradient aggregation, individual gradients are randomly grouped and concatenated, often mixing unrelated failure modes and producing prompts that fail to generalize. To address these limitations, we propose AgentGrad, a prompt optimization framework for multi-agent systems based on sequential intervention and semantic textual gradient abstraction. For each failure, sequential intervention injects a hint into one agent at a time to identify the target agent whose modification resolves the failure. The modified output of the target agent then serves as agent-level supervision for extracting a fine-grained gradient. Semantic textual gradient abstraction clusters semantically similar gradients to prevent mixing unrelated failure modes, and abstracts each cluster into a generalized gradient that captures the shared corrective pattern. Experimental results show that AgentGrad achieves state-of-the-art performance across five MAS benchmarks while reducing wall-clock optimization time by 2.5 times and optimization cost by 21.8% on average compared to the next-best baselines. Our code is available at https://github.com/mlvlab/AgentGrad.