Abstract
As AI services, cloud platforms, and data infrastructures become deeply embedded in enterprise service operations, intelligent customer service systems are no longer limited to front-end response tools. In many firms, they now function as internal workbenches that support scheduling, operational monitoring, quality inspection, training, and knowledge management in addition to customer-facing interactions. Prior work has largely emphasized algorithmic performance, automation rates, or customer satisfaction, while paying much less attention to the employee experience of internal users who work across complex service processes. This study examines an enterprise intelligent customer service workbench from that perspective. We adopt a research path that combines grounded-theory-based inquiry, design translation, and formative usability evaluation. First, we conducted 13 semi-structured interviews with frontline agents, team leaders, QA specialists, trainers, supervisors, and operations managers. Through open, axial, and selective coding, we identified recurring experience problems across roles. Second, we translated these findings into three design lines: tool efficiency, information ecology, and collaboration support. These lines were then instantiated in three prototype modules for scheduling management, operations insight, and knowledge support. Third, we invited 131 participants to complete unified task scenarios and evaluated the prototypes with a module-adapted 10-item usability instrument derived from the System Usability Scale (SUS) response structure. The findings show that the core problems of the existing workbench can be traced to fragmented task chains, weak information interpretability, and poor mechanisms for routing knowledge and collaboration outcomes back into the system. The formative evaluation reports average 0-100 usability scores of 81.47 for scheduling management, 78.42 for operations insight, and 78.00 for knowledge support, with no statistically significant difference across user groups (p > 0.05). By shifting attention from consumer-facing service encounters to the employee experience of internal service work, this paper adds contextual evidence to intelligent customer service research and offers a reusable evidence-to-design pathway for complex B2B workbenches.
Recommended Citation
Sun, Kaiqiang; He, Guangyu; Liao, Rui; Li, Weixi; and Liu, Lei, "Improving Multi-Role Intelligent Customer Service Workbenches in Enterprise Settings: A Grounded-Theory-Informed Design and Formative Evaluation" (2026). ICEB 2026 Proceedings (Beijing, China). 50.
https://aisel.aisnet.org/iceb2026/50