@inproceedings{chang-etal-2026-flowswitch,
    title = "{F}low{S}witch: A State-Aware Framework for Workflow Transitions in Adaptive Dialogue Agents",
    author = "Chang, Wen Yu  and
      Qiu, Luning  and
      Liu, Yi-Hung  and
      Chen, Yun-Nung",
    editor = "Riccardi, Giuseppe  and
      Mousavi, Seyed Mahed  and
      Torres, Maria Ines  and
      Yoshino, Koichiro  and
      Callejas, Zoraida  and
      Chowdhury, Shammur Absar  and
      Chen, Yun-Nung  and
      Bechet, Frederic  and
      Gustafson, Joakim  and
      Damnati, G{\'e}raldine  and
      Papangelis, Alex  and
      D{'}Haro, Luis Fernando  and
      Mendon{\c{c}}a, John  and
      Bernardi, Raffaella  and
      Hakkani-Tur, Dilek  and
      Di Fabbrizio, Giuseppe {''}Pino{''}  and
      Kawahara, Tatsuya  and
      Alam, Firoj  and
      Tur, Gokhan  and
      Johnston, Michael",
    booktitle = "Proceedings of the 16th International Workshop on Spoken Dialogue System Technology",
    month = feb,
    year = "2026",
    address = "Trento, Italy",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/dashboard-stats/2026.iwsds-1.2/",
    pages = "18--33",
    abstract = "To enhance large language models ({LLM}s) with real-world task-solving capabilities, integrating workflow knowledge into {LLM}s has emerged as a promising direction. However, real-world conversations are inherently dynamic{---}users often shift intents or request actions beyond the scope of the current workflow. Existing systems struggle to detect such transitions and to decide when to retrieve or switch to a new workflow. This paper presents {F}low{S}witch, a state-aware framework that learns when to search for relevant workflows and switch between them during multi-turn dialogues. A policy module determines whether to continue within the current workflow or transition to a new one based on contextual representations. When searching, a retriever identifies the most relevant workflow knowledge given the dialogue state. We conduct comprehensive experiments to explore the optimal configuration of {F}low{S}witch, including workflow format, retrieval input type, and retrieval method. Experimental results show that our framework, when using the agent{'}s self-generated search queries, achieves the highest Top-1 accuracy and Mean Average Precision ({MAP}). Moreover, {F}low{S}witch reduces nearly 50{\%} of search operations, substantially lowering computational cost and response time."
}