Intelligent conversational agents (ICAs) are revolutionizing how humans interact with information systems. Designed to provide human-like service, ICAs are generally evaluated by users in comparison to their human counterparts, often resulting in less-than-expected user experiences. Our research investigates user acceptance of ICAs in this suboptimal condition of commercial customer service. Drawing from the dual perspectives of expectancy confirmation theory and task technology fit theory, we theorize and test an integrated research model on the collective impact of user expectancy confirmation regarding ICA capabilities and their assessment of service-ICA fit on user acceptance. Results from a field survey of 350 users of five ICAs deployed by major Canadian telecom service providers reveal the significant influence of both user expectancy confirmation with ICA capabilities and their assessment of ICA fit-to-service, with the latter playing a more prominent role in shaping user acceptance. Even though ICA performance may not always meet user expectations, users are still willing to engage with ICA services when they perceive the ICA as a fitting solution for their specific service complexity and availability requirements.

中文翻译：

---

低于预期但仍愿意使用：用户对当前智能会话代理的接受度


智能会话代理 (ICA) 正在彻底改变人类与信息系统交互的方式。 ICA 旨在提供类似人类的服务，用户通常会与人类同行进行比较来评估 ICA，这通常会导致用户体验低于预期。我们的研究调查了在商业客户服务不理想的情况下用户对 ICA 的接受程度。从期望确认理论和任务技术契合理论的双重角度出发，我们理论化并测试了一个综合研究模型，该模型涉及用户期望确认对 ICA 能力的集体影响以及服务 ICA 契合度对用户接受度的评估。对加拿大主要电信服务提供商部署的 5 种 ICA 的 350 名用户进行的现场调查结果显示，用户对 ICA 功能的预期确认以及对 ICA 服务适合度的评估都具有显着影响，其中后者在影响因素中发挥着更为突出的作用。塑造用户接受度。尽管 ICA 性能可能并不总是满足用户期望，但当用户认为 ICA 适合其特定服务复杂性和可用性要求时，他们仍然愿意使用 ICA 服务。