Abstract
In the era of rapid advancements in artificial intelligence (AI), AI tools are fundamentally transforming the ways in which people learn and work. However, their application within real-world organizational settings remains in an exploratory stage—particularly in the large-scale deployment of AI assistants tailored to specific business contexts.Can AI assistants effectively support human work? What impact do they have on organizational performance, and what are the underlying mechanisms of this impact? These questions remain largely unexplored in existing research. This study addresses this gap by examining the role of AI assistants in a highly complex enterprise sales environment. Specifically, it investigates how AI assistants influence organizational performance and explores both their design principles and the mechanisms through which they enhance performance.
Through a field experiment using real-world data from 336 sales personnel, we found that the group utilizing AI assistants experienced a 12% overall increase in order volume, indicating a significant improvement in organizational performance. Moreover, we discovered that this performance gain was not driven by an increase in workload, but rather by improvements in work quality facilitated by the transfer of tacit knowledge. Specifically, AI assistants enhanced communication efficiency, thereby reducing the working hours required of sales experts. Additionally, by improving demand assessment and the judgment of purchase intentions, AI assistants optimized time efficiency and the utilization of communication resources, ultimately leading to overall performance growth.
Furthermore, we also observed heterogeneous effects based on the capability levels of sales personnel. For high-performing individuals, AI assistants primarily enhanced time efficiency. For average performers, improvements were observed in both time efficiency and resource allocation. Even lower-performing employees benefited significantly, although their performance gains required greater effort due to limitations in internalizing tacit knowledge.
Grounded in the theory of tacit knowledge transfer and informed by the operational principles of AI, this research develops an AI-driven extension of the "SECI" model. It explains how tacit knowledge is transferred within organizations through AI and why this process improves organizational performance.
This study offers high practical value. It provides actionable insights into how to leverage AI assistants to enhance performance in AI-integrated environments and presents guiding principles for designing AI assistants and continuously improving their contribution to organizational outcomes.
| Date of Award | 18 Mar 2025 |
|---|---|
| Original language | Chinese (Simplified) |
| Awarding Institution |
|
| Supervisor | Yu Zhang (Supervisor) & Yan Gong (Supervisor) |
Keywords
- AI
- AI Assistant
- Organizational Performance
- Tacit Knowledge
- Field Experiment
- Knowledge Management
- SECI Model