联系我们

论文查询结果	返回搜索

论文编号：	16042
作者编号：	2120243855
上传时间：	2026/6/3 17:57:21
中文题目：	售前咨询场景中AI客服语言确定性对消费者联系方式披露的影响研究
英文题目：	Research on the Impact of AI Customer Service Language Certainty on Consumer Contact Information Disclosure in Pre-Sales Consultation Contexts
指导老师：	任星耀
中文关键字：	AI客服；语言确定性；联系方式披露；语言功能；客户旅程阶段
英文关键字：	AI Customer Service; Language Certainty; Contact Information Disclosure; Language Function; Customer Journey Stage
中文摘要：	在智能化营销环境中，AI客服正逐渐成为售前咨询中的重要触点。相较于传统人工客服，AI客服不仅能够规模化响应消费者咨询，也使语言表达更具可配置性。在售前获客场景中，联系方式披露是一种兼具隐私风险、关系推进与后续触达价值的高敏感行为，因此，AI客服“如何说”成为影响消费者是否愿意进入后续关系的重要因素。尽管已有研究指出语言确定性具有双刃剑特征，但其在真实AI客服售前咨询中的行为后果，尤其是在不同语言功能和客户旅程阶段下的差异性影响，仍缺乏探讨。 本文以某教育服务平台的真实售前咨询对话日志为研究对象，以联系方式披露为因变量，基于文本分析方法识别AI客服语句的语言确定性及其语言功能，并结合客户旅程阶段标签，检验不同类型语言确定性的差异化作用及其边界条件；同时，通过更换测量方式、更换估计模型和扩展样本范围进行稳健性检验。 研究发现：第一，AI客服信息导向语言确定性与行动导向语言确定性对消费者联系方式披露具有方向相反的影响。信息导向语言确定性越高，消费者越可能披露联系方式；行动导向语言确定性越高，消费者越不可能披露联系方式。第二，客户旅程阶段主要调节信息导向语言确定性的作用，其正向作用主要体现在信息搜寻阶段，且显著强于需求识别阶段。第三，客户旅程阶段未显著调节行动导向语言确定性的负向影响，其抑制作用在不同购前阶段中相对稳定。 本文从语言表达层面推进了AI客服研究，揭示了语言确定性的作用具有功能依赖性与阶段依赖性，并为解释既有研究结论的不一致提供了更具情境性的视角。在实践上，本文也为售前咨询场景中AI客服的语言设计、效果评估与阶段化运营提供了一定参考。本文表明，售前AI客服不宜采用统一强度的高确定性或低确定性表达，而应依据语言功能与客户旅程阶段进行差异化配置：在信息支持型表达中可适度提高确定性，在行动推进型表达中则应避免过强、过早的确定性，以提升售前咨询中的线索转化效果。
英文摘要：	In the intelligent marketing environment, AI customer service has gradually become an important touchpoint in pre-sales consultations. Compared with traditional human customer service, AI customer service not only enables scalable responses to consumer inquiries but also makes language expression more configurable. In pre-sales lead generation contexts, contact information disclosure is a highly sensitive behavior involving privacy risk, relationship advancement, and the value of subsequent contact. Accordingly, how AI customer service communicates has become an important factor influencing whether consumers are willing to enter a subsequent relationship. Although prior research has suggested that language certainty has a double-edged nature, its behavioral consequences in real pre-sales AI customer service interactions, especially its differential effects across language functions and customer journey stages, remain underexplored. Using real pre-sales consultation dialogue logs from an education service platform, this thesis takes contact information disclosure as the dependent variable. Based on text analysis, it identifies the language certainty and language functions of AI customer service utterances and, in combination with customer journey stage classifications, examines the differential effects of different types of language certainty and their boundary conditions. In addition, robustness tests are conducted by adopting alternative measurements, alternative estimation models, and expanded sample scopes. The findings are as follows. First, informational language certainty and action-oriented language certainty exert opposite effects on consumers’ contact information disclosure. The higher the informational language certainty, the more likely consumers are to disclose their contact information; the higher the action-oriented language certainty, the less likely they are to do so. Second, customer journey stage mainly moderates the effect of informational language certainty: its positive effect is primarily manifested in the information search stage and is significantly stronger than in the need recognition stage. Third, customer journey stage does not significantly moderate the negative effect of action-oriented language certainty; its inhibitory effect remains relatively stable across different pre-purchase stages. This thesis advances research on AI customer service from the perspective of language expression by showing that the effect of language certainty is both function-dependent and stage-dependent, and by offering a more contextualized explanation for inconsistencies in prior findings. Practically, the thesis also provides implications for the language design, performance evaluation, and stage-based operation of AI customer service in pre-sales consultation settings. The findings suggest that pre-sales AI customer service should not adopt uniformly high-certainty or low-certainty expressions. Instead, language certainty should be configured differentially according to language function and customer journey stage: certainty may be moderately increased in informational support-oriented expressions, whereas excessively strong or premature certainty should be avoided in action-promotion-oriented expressions so as to improve lead conversion in pre-sales consultations.
查看全文：	预览  下载(下载需要进行登录)