Online reviews help people make better decisions. Review platforms usually depend on typed input, where leaving a good review requires significant effort because users must carefully organize and articulate their thoughts. This may discourage users from leaving comprehensive and high-quality reviews, especially when they are on the go. To address this challenge, we developed Vocalizer, a mobile application that enables users to provide reviews through voice input, with enhancements from a large language model (LLM). In a longitudinal study, we analysed user interactions with the app, focusing on AI-driven features that help refine and improve reviews. Our findings show that users frequently utilized the AI agent to add more detailed information to their reviews. We also show how interactive AI features can improve users self-efficacy and willingness to share reviews online. Finally, we discuss the opportunities and challenges of integrating AI assistance into review-writing systems.
Online reviews have played a pivotal role in consumers' decision-making processes. Existing research has highlighted the significant impact of managerial review responses on customer relationship management and firm performance. However, a large portion of online reviews remains unaddressed due to the considerable human labor required to respond to the rapid growth of online reviews. While generative AI has achieved remarkable success in a range of tasks, they are general-purpose models and may not align well with domain-specific human preferences. To tailor these general generative AI models to domain-specific applications, finetuning is commonly employed. Nevertheless, several challenges persist in finetuning with domain-specific data, including hallucinations, difficulty in representing domain-specific human preferences, and over conservatism in offline policy optimization. To address these challenges, we propose a novel preference finetuning method to align an LLM with domain-specific human preferences for generating online review responses. Specifically, we first identify the source of hallucination and propose an effective context augmentation approach to mitigate the LLM hallucination. To represent human preferences, we propose a novel theory-driven preference finetuning approach that automatically constructs human preference pairs in the online review domain. Additionally, we propose a curriculum learning approach to further enhance preference finetuning. To overcome the challenge of over conservatism in existing offline preference finetuning method, we propose a novel density estimation-based support constraint method to relax the conservatism, and we mathematically prove its superior theoretical guarantees. Extensive evaluations substantiate the superiority of our proposed preference finetuning method.
Voice agents increasingly require reliable tool use from speech, whereas prominent tool-calling benchmarks remain text-based. We study whether verified text benchmarks can be converted into controlled audio-based tool calling evaluations without re-annotating the tool schema and gold labels. Our dataset-agnostic framework uses text-to-speech, speaker variation, and environmental noise to create paired text-audio instances while preserving the original dataset annotations. Based on extensive evaluation of 7 omni-modal models on audio-converted versions of Confetti and When2Call, our framework demonstrates that the performance is strongly model- and task-dependent: Gemini-3.1-Flash-Live obtains the highest Confetti score (70.4), whereas GPT-Realtime-1.5 performs best on When2Call (71.9). On Confetti, the text-to-voice gap ranges from 1.8 points for Qwen3-Omni to 4.8 points for GPT-Realtime-1.5. A targeted analysis of failure cases demonstrates that degradations most often reflect misunderstandings of argument values in the speech. Considering real-world deployment scenarios, we further report text-only results, an ambiguity-based reformulation stress test, and a reference-free LLM-as-judge protocol validated against human preferences. Notably, we find that open-source Qwen3 judges with at least 8B parameters exceed 80% agreement with proprietary judges, supporting privacy-preserving evaluation. Overall, our framework provides a verifiable and reproducible first-stage diagnostic that complements purpose-built audio corpora.
Evaluating conversational voice agents at scale re- quires reliable assessment methods that capture both observ- able interaction quality and the contextual judgment typically provided by human evaluators. We investigate LLM-as-a-Judge evaluation by comparing human judgments with GPT-4.1 and GPT-5 on telecom and retail voice-agent conversations, across conversational quality and safety dimensions. The same interac- tions are scored under three evaluation configurations, p0, p1, and p2, to test whether automated judgments are sensitive to the evaluation setup and whether observed patterns generalize across configurations and judge models. Beyond aggregate agreement, we examine metric-level correlations, evaluator consistency, and systematic human-LLM disagreement to identify which conver- sational attributes can be judged reliably by automation and which remain sensitive to interpretation and context. Effective voice-agent evaluation is also shaped by pipeline-level factors such as speech generation, streaming, and error propagation across ASR, reasoning, and tool-calling stages, motivating our focus on comparing how human and LLM judges score the same interactions end to end. Our results show that LLM- based evaluation can serve as an effective component of large- scale voice-agent assessment, but that its reliability is metric- and configuration-dependent rather than uniform. This pro- vides an empirical framework for identifying which metrics suit automated evaluation and supports hybrid pipelines in which LLM judges handle scalable assessment while human evaluators remain engaged for metrics that demand contextual interpretation and higher-confidence judgment.