Abstract:In this paper, we present the technical details and periodic findings of our project, CareerAgent, which aims to build a generative simulation framework for a Holacracy organization using Large Language Model-based Autonomous Agents. Specifically, the simulation framework includes three phases: construction, execution, and evaluation, and it incorporates basic characteristics of individuals, organizations, tasks, and meetings. Through our simulation, we obtained several interesting findings. At the organizational level, an increase in the average values of management competence and functional competence can reduce overall members' stress levels, but it negatively impacts deeper organizational performance measures such as average task completion. At the individual level, both competences can improve members' work performance. From the analysis of social networks, we found that highly competent members selectively participate in certain tasks and take on more responsibilities. Over time, small sub-communities form around these highly competent members within the holacracy. These findings contribute theoretically to the study of organizational science and provide practical insights for managers to understand the organization dynamics.
What problem does this paper attempt to address?
The problem this paper attempts to address is: constructing a generative organizational behavior simulation framework based on large language models (LLM) to study the impact of employee capabilities on work stress, task completion, and circle division under the Holacracy management model. Specifically, the paper aims to:
1. **Analyze the impact of employee capabilities at the organizational level**: Explore how average management and functional capabilities affect the overall stress level, task completion, and circle formation of organizational members.
2. **Analyze the impact of employee capabilities at the individual level**: Study how different levels of individual capabilities affect personal work stress, task completion, workload, and the formation of social circles.
### Background and Motivation
Holacracy is an innovative management model proposed by software company founder Brian Robertson, aiming to achieve decentralized organizational management through role distribution. Although this model can enhance employee creativity, it may also lead to conflicts between roles and teams, causing many organizational practices to fail. While some static impact mechanisms have been explored in the past, processes such as delegation in dynamic system operations have not been fully understood. Therefore, this paper utilizes the simulation capabilities of large language models (LLM) to construct an organizational behavior simulation framework named CareerAgent to deeply study organizational behavior under the Holacracy management model.
### Main Findings
- **Organizational Level**:
- Improving average management and functional capabilities can reduce the overall stress level of members but negatively impact deep performance indicators of the organization, such as the average task completion rate.
- The number of circles and the number of people in each circle decrease with the improvement of average capabilities, indicating that high-capability members tend to participate in fewer circles but take on more responsibilities.
- **Individual Level**:
- Improving management and functional capabilities can enhance members' work performance.
- High-capability members selectively participate in certain tasks and take on more responsibilities, forming small sub-communities around them over time.
### Experimental Design
The experiment is based on different levels of organizational average capabilities, including two categories of high and low management and functional capabilities, and three categories of variability (high, medium, low) in management and functional capabilities, totaling 36 experimental groups. Each experiment lasts for eight weeks, with task allocation and selection each week, and each round includes four tasks.
### Results Analysis
- **Organizational Level**:
- Improving management and functional capabilities significantly reduces the overall stress level of members (management capability: β1=-8.417, p < 0.01; functional capability: β2=-8.099, p < 0.01).
- However, the improvement of these two capabilities has a negative effect on overall task completion (management capability: β1=-0.080, p < 0.05; functional capability: β2=-0.084, p < 0.05), meaning that the higher the average capability of organizational members, the worse the task completion.
- **Individual Level**:
- Improving individual management and functional capabilities can enhance task completion and reduce work stress.
- High-capability members tend to participate in fewer circles but take on more workload.
### Conclusion
This study provides new insights into organizational science from both theoretical and practical perspectives, helping managers better understand the organizational dynamics under the Holacracy management model and providing guidance for practical applications.