ECLAIR: Enhanced Clarification for Interactive Responses
John Murzaku, Zifan Liu, Md Mehrab Tanjim, Vaishnavi Muppala, Xiang Chen, Yunyao Li
TL;DR
ECLAIR introduces a unified, end-to-end framework for interactive disambiguation in enterprise AI assistants by integrating multiple ambiguity-detection and grounding agents into a single prompt, enabling simultaneous ambiguity decision and clarifying question generation. Evaluated on real Adobe Experience Platform data, ECLAIR achieves higher precision in deciding when to ask clarifications and produces clarifications more aligned with human-annotated gold questions than standard few-shot baselines, while maintaining practicality in production settings. The approach demonstrates two production use cases with UI-driven disambiguation, and discusses deployment considerations including precision targets, latency management, user studies, and continuous monitoring. Overall, ECLAIR offers a modular, agent-based architecture that enhances context-awareness and flexibility for enterprise AI assistants, providing a practical blueprint for deploying robust disambiguation in real-world systems.
Abstract
We present ECLAIR (Enhanced CLArification for Interactive Responses), a novel unified and end-to-end framework for interactive disambiguation in enterprise AI assistants. ECLAIR generates clarification questions for ambiguous user queries and resolves ambiguity based on the user's response.We introduce a generalized architecture capable of integrating ambiguity information from multiple downstream agents, enhancing context-awareness in resolving ambiguities and allowing enterprise specific definition of agents. We further define agents within our system that provide domain-specific grounding information. We conduct experiments comparing ECLAIR to few-shot prompting techniques and demonstrate ECLAIR's superior performance in clarification question generation and ambiguity resolution.