@inproceedings{srinivas-etal-2026-thinking,
title = "Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice Agents",
author = "Srinivas, Vidya and
Englhardt, Zachary and
Iyer, Vikram and
Patel, Shwetak",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1371/",
doi = "10.18653/2026.findings-emnlp.1371",
pages = "29148--29187",
ISBN = "979-8-89176-462-0",
abstract = "Voice agents face a fundamental tension: the reasoning, retrieval, and tool use that make foundation models capable are iterative and slow, while conversational interaction demands prompt responses. Smaller, real-time models meet the latency bar but do not match foundation models on complex tasks, leaving current voice agents to trade off responsiveness and capability. We introduce conversational infill, where a small Talker model both fluently integrates streamed Reasoner model knowledge into its responses during inference and immediately generates contextually grounded responses to hide external Reasoner latency. We curate a 290,571-example dataset and demonstrate that this task is learnable across seven small language models (135M to 1.7B parameters).\par Our system, ConvFill, runs on an Apple M2 with millisecond-level latency, while remaining within 6.3{\%} of corresponding frontier Reasoner performance. Conversational infill unlocks a new point on the latency{--}capability Pareto frontier, offering a practical path toward responsive and capable voice agents."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="srinivas-etal-2026-thinking">
<titleInfo>
<title>Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vidya</namePart>
<namePart type="family">Srinivas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zachary</namePart>
<namePart type="family">Englhardt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vikram</namePart>
<namePart type="family">Iyer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shwetak</namePart>
<namePart type="family">Patel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Voice agents face a fundamental tension: the reasoning, retrieval, and tool use that make foundation models capable are iterative and slow, while conversational interaction demands prompt responses. Smaller, real-time models meet the latency bar but do not match foundation models on complex tasks, leaving current voice agents to trade off responsiveness and capability. We introduce conversational infill, where a small Talker model both fluently integrates streamed Reasoner model knowledge into its responses during inference and immediately generates contextually grounded responses to hide external Reasoner latency. We curate a 290,571-example dataset and demonstrate that this task is learnable across seven small language models (135M to 1.7B parameters).\par Our system, ConvFill, runs on an Apple M2 with millisecond-level latency, while remaining within 6.3% of corresponding frontier Reasoner performance. Conversational infill unlocks a new point on the latency–capability Pareto frontier, offering a practical path toward responsive and capable voice agents.</abstract>
<identifier type="citekey">srinivas-etal-2026-thinking</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1371</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1371/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>29148</start>
<end>29187</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice Agents
%A Srinivas, Vidya
%A Englhardt, Zachary
%A Iyer, Vikram
%A Patel, Shwetak
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F srinivas-etal-2026-thinking
%X Voice agents face a fundamental tension: the reasoning, retrieval, and tool use that make foundation models capable are iterative and slow, while conversational interaction demands prompt responses. Smaller, real-time models meet the latency bar but do not match foundation models on complex tasks, leaving current voice agents to trade off responsiveness and capability. We introduce conversational infill, where a small Talker model both fluently integrates streamed Reasoner model knowledge into its responses during inference and immediately generates contextually grounded responses to hide external Reasoner latency. We curate a 290,571-example dataset and demonstrate that this task is learnable across seven small language models (135M to 1.7B parameters).\par Our system, ConvFill, runs on an Apple M2 with millisecond-level latency, while remaining within 6.3% of corresponding frontier Reasoner performance. Conversational infill unlocks a new point on the latency–capability Pareto frontier, offering a practical path toward responsive and capable voice agents.
%R 10.18653/2026.findings-emnlp.1371
%U https://aclanthology.org/2026.findings-emnlp.1371/
%U https://doi.org/10.18653/2026.findings-emnlp.1371
%P 29148-29187
Markdown (Informal)
[Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice Agents](https://aclanthology.org/2026.findings-emnlp.1371/) (Srinivas et al., Findings 2026)
ACL