Towards a Taxonomy for Autonomy in Large-Scale Agile Software Development

2025-03-04

AI Summary

In large-scale agile development, the dynamic balance between team autonomy and organizational governance lacks a systematic, theory-grounded framework. Method: This paper introduces the first taxonomy for autonomy in large-scale agile contexts, proposing a novel five-level hierarchical model and delineating 21 autonomy categories across technical, social, and organizational dimensions—each with rigorously defined degrees of autonomy and constraint boundaries. The taxonomy was developed through an integrated approach combining empirical research, expert consensus, and case-driven modeling. Contribution/Results: The framework fills a critical theoretical gap in autonomy governance for large-scale agile, and industrial validation demonstrates its effectiveness and practicality in identifying redundant controls and optimizing cross-team authorization mechanisms.

Technology Category

• Multiagent Systems: Agent/AI Theories and Architectures
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Game Design — Virtual Humans, NPCs and Autonomous Characters

Application Category

• Security and Privacy: Large-scale security measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience

Abstract

Agile development relies on self-organizing teams having a high degree of autonomy. For single-team development, more autonomy is generally considered better. In large-scale agile development, where several teams collaborate on the same software with technical and social dependencies, finding the right balance between autonomy and organizational control becomes critical. To this end, it is helpful to have a framework that helps reason about autonomy to help map out to what degree teams can be autonomous, and in what aspects that autonomy needs to be restricted. This paper presents our work towards building a framework for autonomy in large-scale agile development. The preliminary version identifies five levels, and 21 categories of autonomy grouped into three areas. These aspects of autonomy can be taken into account when analyzing or defining the limits of autonomy in large-scale agile software development. The use of the framework is illustrated with an example.

Problem

• Balancing autonomy and control in large-scale agile development.
• Developing a framework to define team autonomy levels.
• Identifying categories of autonomy for multi-team collaboration.

Innovation

• Framework for autonomy in large-scale agile development
• Five levels and 21 categories of autonomy
• Balancing autonomy and organizational control