Journal of Competences, Strategy & Management · 2024
How does AI change the way organizations work, and why can it also make established ways of working more stable?
Mahringer, C. A., Danner-Schröder, A., Müller-Seitz, G., & Renzl, B. (2024). How does artificial intelligence promote change and stability of organizational routines? The role of automation and augmentation. Journal of Competences, Strategy & Management, 12, 1–17. 10.25437/jcsm-vol12-104
Artificial intelligence is often associated with transformation. Organizations adopt AI to automate tasks, improve decisions, generate new ideas, and redesign work. Yet AI can also make established ways of working more persistent by moving actions outside human attention or by making people increasingly dependent on algorithmic outputs. Understanding AI in organizations therefore requires attention to both change and stability.
The paper examines this tension through two prominent ways of using AI. Automation refers to machines taking over actions that were previously performed by humans. Augmentation refers to humans and AI working together. Rather than treating these as alternative strategies for an entire process, the paper examines how automated and augmented actions can coexist and interact within the same pattern of organizational work.
We ask: How does the adoption of artificial intelligence promote change and stability?
This is a conceptual paper based on narrative process theorizing. We analyze automation and augmentation across a range of business examples and use them to develop mechanisms that explain how AI affects patterns of organizational action. This approach shifts attention from asking whether a complete job or process is automated toward examining what happens to particular actions and their interdependencies when AI becomes part of everyday work.
This paper examines the influence of artificial intelligence (AI) on the change and stability of organizations. We focus on automation and augmentation as key dimensions of AI and elaborate their effects on organizational routines. AI can promote change of organizational routines through capacitating new actions or reframing patterns of actions, but also their stability through shielding actions and adhering to actions. Moreover, we suggest that these mechanisms can occur simultaneously and sequentially in different parts of routines. This paper contributes to research on automation and augmentation by explaining how these two applications form a duality. While prior research suggested that actors iterate between both applications over time, we suggest that zones of automation and augmentation coexist within different parts of the action patterns of the same routines. Seen this way, humans and AI work hand in hand to perform those routines. We also contribute to Routine Dynamics research by suggesting mechanisms through which AI may lead to the change and stability of routines.
The paper explains why introducing AI can transform some parts of work while simultaneously reinforcing stability in others.
AI promotes change through capacitating, by enabling actions that were previously impossible or too resource-intensive, and reframing, by helping actors see established patterns of work differently. AI promotes stability through shielding, by keeping actions outside actors’ immediate attention, and adhering, by making deviation from established actions more difficult.
Organizations do not necessarily automate or augment an entire routine. Different actions and relations can form zones of automation and augmentation within the same pattern of work. A hiring process, for example, may automate screening while augmenting assessment and leaving interviews largely human.
The paper conceptualizes automation and augmentation as a duality. Machine actions shape what humans can do next, while human actions provide inputs and conditions for subsequent machine actions. Human and AI contributions therefore mutually constitute how the overall work process is performed.
The paper contributes to research on AI, automation, and augmentation by moving below the level of whole jobs or processes. It shows that automation and augmentation can coexist within different parts of the same workflow and can be mutually constitutive rather than sequential stages on a path toward full automation.
It also connects AI research to organizational change by explaining specific mechanisms through which AI can produce divergent outcomes. The same organization can experience transformation and stabilization at once, depending on which actions AI capacitates or reframes and which actions it shields or makes harder to deviate from.
Managers should avoid evaluating AI implementation only at the level of entire jobs or business processes. A more useful question is which specific actions are automated, which are augmented, and how those actions depend on one another. This helps reveal where AI is creating new possibilities and where it may be locking existing practices in place.
The framework also highlights a governance risk. When AI shields actions from human attention or makes subsequent work dependent on algorithmic outputs, opportunities for reflection and change can decline. Organizations therefore need to examine where human oversight and opportunities to question established patterns should remain part of AI-enabled work.
