An Actionable Diagnosis of Multilingual, Multi-Agent Planning Failures
For developers of multilingual multi-agent systems, this work provides a diagnostic framework and a mitigation that improves performance, especially in low-resource languages.
The paper studies planning failures in multilingual multi-agent systems, deriving a taxonomy of planning-grounding failures and introducing TART, a method that improves performance by making taxonomy aspects explicit. TART raises accuracy by 5.6 percentage points on multilingual GAIA across eleven languages.
Multilingual multi-agent systems exhibit substantial degradation beyond English, yet prior work rarely identifies how task-critical information is lost when user requests are converted into executable plans. We study the planner in a multi-agent system as the request-to-action interface and derive an actionable taxonomy of planning-grounding failures from failed real-world task executions. LLM-based analysis shows that these failures constitute an increasing share of unsuccessful executions as language-resource availability declines, with the strongest effects in low-resource languages. To test whether the taxonomy supports mitigation, we introduce TART, Taxonomy-Guided Actionable Representation, that makes the taxonomy's key aspects explicit to the planner and downstream sub-agents. Across multiple languages, three LLM backbones, two datasets, and two agentic configurations, TART consistently improves performance. On multilingual GAIA, it raises a state-of-the-art system's accuracy by 5.6 percentage points averaged across eleven languages spanning low- to high-resource settings.