Beyond Oracle Communication: Benchmarking Interactive Intent Alignment Under Miscommunication and Evolving User Intent

2026-09-29

AI Summary

This study addresses the challenge of aligning LLM agents with users' true needs when intentions are miscommunicated or goals dynamically shift. To this end, it proposes a novel interactive intent alignment paradigm and constructs Drift-Bench++, a benchmark that simulates limited user patience and intent drift through a controllable misalignment pipeline, multi-agent interaction simulation, and silent conditional intent mutation modeling. Furthermore, the GRIP multidimensional evaluation protocol is introduced to overcome the limitations of conventional static benchmarks. Experimental results demonstrate that while enhanced interactivity improves performance, a substantial gap remains relative to ideal behavior, revealing the prevalence and severity of such failures in real-world deployments.

Abstract

Modern LLM agents increasingly tackle complex tasks through interactive, long-horizon exchanges with users, while existing benchmarks generally assume that users always accurately and sufficiently communicate a fixed intent. However, this oracle communication assumption rarely holds in practice: users may miscommunicate, change their goals, and run out of patience. We define this task setting as Interactive Intent Alignment, where agents must recover and continuously track the user's current intent despite imperfect communication and evolving goals. To study this setting, we introduce Drift-Bench++, a principled benchmark construction pipeline for verified executable tasks with controlled misalignment and intent shifts, along with an interaction protocol featuring finite patience, diverse simulated users, and silent interaction-conditioned shifts. We further develop GRIP, a comprehensive evaluation protocol covering task grounding, user realism, inquiry effectiveness, and adaptation to evolving intent. Across diverse environments, models, and interaction conditions, stronger interaction consistently helps but remains far from oracle performance; Validation on deployed ProdAgent sessions further shows that the modeled failures are prevalent and consequential in deployment. By providing a unified, executable benchmark for interactive intent alignment, Drift-Bench++ offers a foundation for evaluating and advancing agents under realistic communication and evolving intent.

Innovation

• Interactive Intent Alignment
• Drift-Bench++
• GRIP Evaluation Protocol
• Miscommunication
• Evolving User Intent