2.1 Theoretical Underpinnings

This study is anchored in two complementary perspectives: the dynamic capabilities view (DCV) and organizational information processing theory (OIPT). Together, these frameworks explain how firms adopt AI technologies and build supply chain resilience in response to turbulent environments.

The dynamic capabilities framework (Teece et al., 1997) emphasizes that, in volatile markets, firms must develop capabilities to sense environmental changes, seize emerging opportunities, and reconfigure resources to maintain competitiveness. AI adoption enables firms to sense shifts in customer demand, supplier performance, and market dynamics through predictive analytics and real-time monitoring (Dubey et al., 2020; Belhadi et al., 2021). SCR extends this logic by acting as a higher-order dynamic capability. It enables firms to integrate technological, relational, and organizational resources to withstand and recover from disruptions, thereby safeguarding long-term performance (Yu et al., 2019). From a DCV perspective, AI adoption strengthens resilience by providing tools that enlarge agility, visibility, and reconfigurability in supply chain operations.

OIPT (Galbraith, 1973) suggests that organizations must match their information-processing capacity to the level of uncertainty and complexity in their environment. Market dynamism is marked by rapid technological innovations, changing customer demands, and competitive turbulence (Jaworski & Kohli, 1993), which present significant information-processing challenges. Firms must absorb and analyze large volumes of dynamic, rapidly changing data to make informed, timely decisions. AI technologies enhance organizational information-processing capacity by enabling predictive modeling, anomaly detection, and real-time decision support (Attah et al., 2024). By increasing the scope and speed of information handling, AI helps firms synchronize their decision-making with environmental uncertainty, ultimately supporting resilience.

Integrating DCV and OIPT reveals that AI adoption occupies a unique boundary position between capability deployment and information processing. While DCV emphasizes the strategic routines that enable firms to sense and reconfigure resources, OIPT explains how organizations absorb and analyze environmental data to support those routines. AI technologies operationalize this linkage by transforming environmental signals into actionable insights, thereby acting as the micro-foundational mechanism through which dynamic capabilities function under turbulence. Additionally, DCV explains why firms require new capabilities to adapt to environmental turbulence, while OIPT clarifies how AI adoption intensifies the capacity to process information and respond effectively to uncertainty. This dual-theoretical foundation positions AI adoption as both a dynamic capability and an information-processing mechanism, making it central to the development of resilient supply chains and increased firm performance. In this way, OIPT provides the micro-foundational logic for DCV: information-processing routines form the operational layer through which dynamic capabilities are enacted. AI functions as the connective mechanism. Its analytics translate sensing and seizing (DCV) into timely interpretation and action (OIPT), thereby linking environmental signals to resource reconfiguration and performance outcomes.

2.2 Hypotheses Development

3.1 Sample and Data Collection

Data for this study were collected through a structured survey administered to firms operating in the United States and South Korea. These two contexts were intentionally selected because they represent distinct institutional and technological environments: the United States reflects a digitally mature ecosystem with widespread technological infrastructure. At the same time, South Korea is characterized by advanced manufacturing and logistics sectors operating under intense competitive and environmental pressures. Examining both settings enables a comparative assessment of how institutional conditions influence AI adoption, resilience-building, and firm performance.

Consistent with firm-level research practices, managers were asked to respond on behalf of their organizations. Survey data were collected in June 2025 through email distribution using a randomized sampling approach. Korean firms were selected from the Korean Chamber of Commerce database, while U.S. firms were randomly contacted and invited to participate. The survey instrument was first developed in English and reviewed by the research team, including an American professor, to ensure content validity. It was then translated into Korean by bilingual scholars, and both language versions were cross-checked to ensure accuracy and conceptual equivalence.

The final sample consisted of 297 firms, of which 172 were based in the United States (57.9%) and 125 in Korea (42.1%). To capture variation across industries, the sample included both manufacturing firms (32.7%) and service, logistics, and other firms (67.3%). Firm size was distributed across three categories: firms with fewer than 149 employees (32.0%), those with 150–249 employees (20.8%), and firms with over 250 employees (47.2%). In terms of firm age, 14.5% were less than 5 years old, 33.7% were between 6 and 10 years old, and 51.8% had been operating for more than 11 years. The sample also showed varying levels of internationalization: 40.0% of firms reported that less than 25% of their sales came from international markets, 24.6% between 26% and 50%, and 35.4% above 51%.

3.2 Measures and Instrument Development

All constructs in this study were measured using established scales from prior research to ensure validity and comparability with existing literature. A five-point Likert scale (1 = strongly disagree to 5 = strongly agree) was used for all items. Market dynamism was defined as the degree of rapid and unpredictable changes in customer needs, supplier capabilities, and competitor actions (Jaworski & Kohli, 1993; Belhadi et al., 2021) and was measured using three items. AI adoption was defined as the extent to which firms integrate AI technologies into their supply chain operations to enhance decision-making, efficiency, and security (Abou-Foul et al., 2023; Dubey et al., 2020), and was measured using four items. Supply chain resilience was defined as a firm's ability to anticipate, adapt to, and recover from disruptions through strong, collaborative relationships (Yu et al., 2019; Altay et al., 2018; Belhadi et al., 2021) and was measured using four items. Although resilience encompasses agility, visibility, and redundancy, this study adopts a relationally anchored operationalization that is widely validated in the literature (e.g., Yu et al., 2019; Altay et al., 2018). Relational resilience captures the collaborative and trust-based mechanisms through which firms coordinate adaptive responses, a core pathway emphasized in dynamic capabilities research.

Finally, firm performance was defined as a firm’s relative success in operational and financial outcomes relative to competitors (Koufteros et al., 2014; Srinivasan & Swink, 2018) and was measured using four items. Firm performance is modeled as a composite reflective construct integrating operational and financial dimensions. Prior supply chain studies commonly employ combined performance measures when evaluating firm competitiveness (Koufteros et al., 2014; Srinivasan & Swink, 2018). Therefore, aggregating these indicators aligns with theoretical and empirical precedent.

To ensure construct reliability and validity, the survey items were pre-tested with a small sample of managers in both the United States and Korea. Results from the measurement model showed acceptable factor loadings, composite reliability, and average variance extracted (AVE) values, all of which exceeded recommended thresholds, indicating robust measurement properties.

4.1 Outer Assessment

The outer (measurement) model was evaluated for indicator reliability, convergent validity, and discriminant validity. As shown in Table 3

Discriminant validity was examined using the Fornell–Larcker criterion (Table 4

4.2 Common Method Bias

To account for the possibility of inflated relationships due to standard method variance, Harman’s single-factor test was conducted. The results revealed that the first factor explained 43.413% of the variance, below the recommended threshold of 50%. This suggests that common method bias is unlikely to pose a serious threat to the validity of the findings. In addition, the distinct factor structure identified during the confirmatory assessment further supports the conclusion that a single source of measurement bias does not drive the results. However, it must be mentioned that although Harman's single-factor test is widely utilized to detect common method bias, it presents substantial methodological limitations as follows (Fuller et al., 2016):

1. low sensitivity,

2. insufficient alone,

3. assumption problems, and

4. type II error issues.

4.3 Inner Assessment

The structural model results in Table 5

Collectively, these findings indicate that AI adoption and resilience are central mechanisms linking environmental turbulence to firm performance. Notably, resilience exhibited the strongest path coefficient to performance, stressing its pivotal role as a mediator in the model.

4.4 Structural Model

The model’s explanatory power was evaluated using the coefficient of determination ($R^2$), as recommended by Hair et al. (2014). These results are displayed in Table 6

By contrast, the explanatory power for Korean firms was weaker. AI adoption ($R^2 = 0.091$), supply chain resilience ($R^2 = 0.314$), and firm performance ($R^2 = 0.230$) all fall below the levels observed in the U.S. sample. These results suggest that the structural relationships explain considerably more variance for U.S. firms than for Korean firms, indicating possible contextual or institutional differences in how AI adoption and supply chain resilience translate into performance outcomes.

4.5 Goodness-of-Fit

While PLS-SEM does not offer a global goodness-of-fit (GoF) metric, two proxy measures have been developed to evaluate the model’s overall fit (Hair et al., 2014). Model fit is assessed using multiple indices to ensure robustness (see Table 7

4.6 Mediation

The mediating roles of AI adoption and resilience were tested using bootstrapped indirect effects, as shown in Table 8

These findings emphasize the centrality of resilience in transforming environmental turbulence and technological adoption into performance outcomes. While AI adoption directly increased resilience, its contribution to firm performance was primarily indirect, operating through the enhancement of resilience. This aligns with prior studies suggesting that advanced technologies yield performance benefits only when embedded within resilient supply chain structures (Yu et al., 2019; Modgil et al., 2022).

4.7 Multigroup Analysis (MGA)

Multigroup analysis was conducted to examine whether the hypothesized relationships differed significantly between U.S. and Korean firms (Tables 9

Results in Table 10 revealed several notable cross-national differences. The effect of market dynamism on AI adoption was significantly greater in U.S. firms ($\beta = 0.763$, $p < 0.001$) than in Korean firms ($\beta = 0.302$, $p < 0.01$), suggesting that U.S. firms are more likely to respond to turbulence by investing in AI technologies. Similarly, market dynamism had positive effects on resilience and performance in the U.S. sample ($\beta = 0.498$ and $\beta = 0.190$, respectively), but these relationships were not statistically significant in Korea. This indicates that in the Korean context, environmental turbulence does not necessarily translate into stronger resilience or improved performance.

By contrast, the effect of AI adoption on resilience was more substantial among Korean firms ($\beta = 0.573, p < 0.001$) than among U.S. firms ($\beta = 0.339, p < 0.01$). This suggests that once Korean firms adopt AI, the technology plays a critical role in enhancing resilience. Finally, resilience significantly developed performance in both contexts, though the effect was more substantial in the U.S. ($\beta = 0.638$) than in Korea ($\beta = 0.422$). Overall, these findings support H11, showing that the strength and significance of the relationships vary across national contexts. The differences highlight the significance of institutional, cultural, and infrastructural conditions in shaping how firms adopt AI and build resilience in response to market dynamics. The pronounced cross-national differences observed in explanatory power align with institutional theory. U.S. firms operate in a liberal-market institutional environment characterized by decentralized decision structures, mature digital infrastructure, and lower cultural uncertainty avoidance. These factors support the deployment of proactive capabilities, which explains why market dynamism strongly predicts both AI adoption and resilience. In contrast, Korean firms exhibit more hierarchical coordination and higher uncertainty avoidance, which encourages reactive adaptation once AI technologies are in place. Thus, institutional logics shape not only the adoption of AI but also the strength with which dynamic capabilities translate into performance.

4.8 Moderation

The moderation analysis shows that market dynamism alters the effect of key capabilities on firm performance (see Table 11

5.1 Theoretical Implications

This study makes three key theoretical contributions. First, it combines the dynamic capabilities view and organizational information processing theory to explain how firms respond to turbulence. The results show that AI adoption functions not only as a dynamic capability for sensing and reconfiguring but also as an information-processing mechanism that reduces uncertainty. This dual role broadens both theories by showing that digital technologies can simultaneously augment capability development and information-processing capacity. Second, the findings establish supply chain resilience as a key mediating capability. Previous research identified resilience as important for maintaining performance during disruptions (Yu et al., 2019). However, few studies have empirically substantiated its role as the mechanism through which market dynamism and AI adoption influence performance. Our results verify that resilience, rather than AI adoption alone, is the pathway through which environmental turbulence and digital investment lead to better outcomes. This advances resilience research by positioning it at the intersection of environmental conditions, technology adoption, and firm performance. Third, the study contributes to comparative management research by revealing institutional differences that refine both theoretical perspectives. It extends the dynamic capabilities view by demonstrating that the effectiveness of sensing, seizing, and reconfiguring routines enabled by AI depends on institutional and technological maturity, challenging the assumption of universal capability deployment. It also advances organizational information processing theory by introducing AI as a non-human processing mechanism that complements, rather than replaces, managerial cognition in uncertain environments. Integrating these insights, the study proposes a boundary-conditioned synthesis in which AI acts as the operational conduit linking environmental turbulence, information processing, and capability reconfiguration to sustained performance.

The moderation results further refine both DCV and OIPT. They demonstrate that the performance benefits of AI adoption intensify under heightened market dynamism, confirming that dynamic capabilities yield greater returns in uncertain environments. This supports the OIPT view that information-processing needs increase with turbulence and validates DCV’s proposition that capability effectiveness depends on contextual fit. By contrast, the absence of a moderating effect for resilience suggests that structural adaptability contributes steadily to performance regardless of environmental volatility, accentuating its role as a foundational rather than contingent capability.

The cross-national evidence challenges the assumption that dynamic capabilities necessarily yield similar benefits across contexts. In the U.S., firms convert turbulence into opportunity through direct capability deployment. Korean firms, however, rely more heavily on AI as an enabling mechanism that compensates for structural rigidity and cultural constraints. This indicates that digital technologies can serve as substitutes for specific organizational capabilities in environments where proactive sensing and reconfiguration routines are less developed.

5.2 Managerial and Practical Implications

The findings offer several practical insights for managers and policymakers. For managers in U.S. firms, the results suggest that investing in AI technologies is a natural extension of responding to market dynamism. Because U.S. firms already operate in a digitally mature environment, they can leverage turbulence as an opportunity to strengthen resilience and outperform competitors. For managers in Korean firms, however, the results suggest that AI adoption is particularly crucial for building resilience. In contexts where market dynamism does not automatically drive greater resilience, technology adoption is a key enabler of supply chain continuity.

The moderation results also provide actionable insights for managers. When market conditions are volatile, AI adoption becomes particularly crucial for maintaining performance. Firms should not interpret environmental turbulence solely as a threat but also as an opportunity to extract greater value from their digital investments. Conversely, resilience-building should remain a consistent managerial priority, since its contribution to performance appears stable across environmental conditions. Managers can thus align technological investments with levels of environmental dynamism to achieve optimal performance gains.

For both contexts, the study highlights the importance of resilience as a strategic capability. Managers should not assume that AI adoption alone will lead to performance improvements; instead, expertise must be embedded in resilience-building practices, such as collaborative supplier relationships, scenario-based planning, and flexible resource allocation. Policymakers can also draw on these findings to design incentives for AI adoption in supply chains, particularly in environments where competitive intensity is high but resilience practices are underdeveloped. To translate these insights into actionable strategies, managers should view AI not as a stand-alone technological tool but as an integral element of their capability-building processes. In practice, this requires embedding AI within cross-functional decision routines so that the information it generates directly informs sensing, seizing, and reconfiguring activities.

Managers can strengthen this linkage by integrating AI-based analytics with resilience exercises, ensuring that technological insights are continuously transformed into adaptive organizational responses. Effective AI adoption also depends on fostering collaboration and data sharing with suppliers, logistics partners, and other stakeholders, thereby translating algorithmic predictions into coordinated action across the supply network. By cultivating these organizational routines, firms can move beyond using AI merely for operational efficiency and instead leverage it as a strategic enabler of resilience, agility, and sustained performance in turbulent environments.

To operationalize AI as a dynamic capability, managers should embed AI-based analytics into core sensing routines, integrate predictive insights into cross-functional decision cycles, and couple AI outputs with scenario-planning exercises. Managers must also invest in organizational structures that convert information-processing gains into reconfigurable action, such as flexible resourcing, modular processes, and collaborative supplier arrangements.