Next Article in Journal
Research on an Innovation Opportunity Identification Method Based on Link Prediction in Heterogeneous Networks
Previous Article in Journal
How Stablecoins Reshape Interactions Within a Fragmented USD-Access System: Evidence from a Time-Frequency TVP-VAR Connectedness Model
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Systems
Volume 14
Issue 8
10.3390/systems14080933
systems-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

From Technological Enablement to Value Co-Creation: How AI Capability Is Linked to Business Model Innovation in Digital Firms

by
Jiayi Xin
and
Zhen Zhang
*
School of Business, Ningbo University, Ningbo 315211, China
*
Author to whom correspondence should be addressed.
Systems 2026, 14(8), 933; https://doi.org/10.3390/systems14080933
Submission received: 11 June 2026 / Revised: 25 July 2026 / Accepted: 30 July 2026 / Published: 2 August 2026
(This article belongs to the Section Artificial Intelligence and Digital Systems Engineering)
Browse Figure
Versions Notes

Highlights

Please indicate how your work links to systems science via your contributions to systems practice, theory, and/or methodology.
  • This study operationalizes service-dominant logic within a testable nomological network, demonstrating how AI capability, as an operant resource, integrates with organizational routines and institutional norms to drive systemic business model reconfiguration.
  • By integrating PLS-SEM with a three-wave longitudinal design, this research provides a methodological framework for empirically examining complex mediated–moderated relationships in socio-technical systems.
What are the main findings and/or the implications of the main findings?
  • This study advances systems-theoretic understanding of AI-enabled value co-creation by operationalizing service-dominant logic’s core axioms, operant resources and institutional arrangements, within a testable nomological network, demonstrating how AI capability as a higher-order operant resource integrates with organizational routines (customer responsiveness) and institutional norms (digital organizational culture) to drive systemic business model reconfiguration.
  • For managers, translating AI investments into strategic business model innovation requires not only building AI infrastructure and capabilities but also cultivating a digital organizational culture that supports data-driven collaboration, cross-functional experimentation, and evidence-based decision-making to unlock the full value-creation potential of AI.

Abstract

Despite substantial AI technology investment, many firms fail to translate isolated AI applications into integrated capabilities that deliver strategic returns and drive business model changes. Grounded in service-dominant logic (SDL), this study proposes and empirically tests a theoretical framework that positions AI capability (AIC) as a key antecedent in the nomological network of business model innovation (BMI). Drawing on a three-wave, two-week-interval longitudinal survey of 193 Chinese digital-intensive firms across IT, technical services, and digital leasing industries, and employing PLS-SEM, we examine associations among focal constructs, specifically, the mediating role of customer responsiveness (CR) and the moderating effect of digital organizational culture (DOC). This design mitigates common method bias and establishes temporal causal ordering. Empirical results indicate that AIC positively relates to BMI both directly and indirectly through CR, and that DOC significantly enhances the indirect effect of AIC on BMI via CR, particularly under high levels of AI-enabled sensing and interpretation. However, causal inference is limited by the cross-sectional nature of the data and self-reported measures. This study makes three key theoretical contributions. First, we identify CR as a market-oriented mechanism linking AIC to BMI, shifting focus from prior internal efficiency-focused mechanisms to customer-centric value co-creation. Second, we extend SDL to the AI context by clarifying how DOC shapes the strategic transformation of ambiguous probabilistic AI outputs into market-oriented actions. Third, we introduce DOC as an internal boundary condition for AIC, complementing prior research on external environmental moderators. These findings provide actionable guidance for managers seeking to unlock the strategic value of AI investments. Findings reflect statistical associations rather than confirmed causal effects, and results are based on perceptual survey data from Chinese digital firms.

1. Introduction

Business model innovation (BMI) is recognized as a critical source of competitive advantage, enabling organizations to redefine how they create, deliver, and capture value in changing markets [1,2,3]. AI reshapes products, services, and the nature of innovation processes, making it a critical influencing factor of BMI [4,5,6]. While firms acknowledge AI’s strategic potential and deploy AI models to enable value-creation mechanisms, such efforts often fail to deliver substantial progress. In response, a growing body of research has shifted attention from the mere adoption of AI technologies toward understanding how AI capability (AIC) influences BMI. AIC refers to the digital organizational capacity to integrate AI infrastructure, data, and expertise for operational transformation [7,8].
Recent research has empirically established the positive direct association between AIC and BMI [9]. This capability enables firms to reconfigure the core architectures of their business models, which lies at the core of BMI [2]. Along this line, a few studies have begun to elucidate this connection through internally focused mechanisms such as strategic agility [6], organizational adaptability [10] and dynamic capabilities for circular BMI [11]. These studies have offered valuable insights. Nevertheless, two critical research gaps remain under-explored in the current literature. These two gaps are inherently interconnected: even if AI capability can theoretically improve customer responsiveness to transform business models, this transformation process is highly dependent on firms’ internal cultural norms. Without a supportive digital organizational culture, firms cannot effectively interpret ambiguous AI predictions and translate them into customer-centric market actions, which further weakens the indirect path from AIC to BMI via CR. This dual omission jointly limits the completeness of the AIC-BMI nomological network in the existing literature.
First, while existing studies on AIC and BMI have predominantly focused on internal efficiency-oriented mechanisms such as resource orchestration and operational optimization [9], this narrow focus overlooks the market-facing nature of BMI, which fundamentally involves reshaping firm–customer value exchange systems. Although a few studies have touched on response speed [6], there is a lack of empirical research that unpacks how AIC is translated into interactive firm–customer actions that drive systemic business model reconfiguration. Second, prior research on the boundary conditions of AIC has focused almost exclusively on external environmental factors such as environmental dynamism [12]. However, unlike traditional deterministic technologies, AIC generates probabilistic, data-driven insights that require organizational interpretation and collective action. This means that even firms with identical AIC may achieve vastly different strategic outcomes, depending on their internal shared norms and cultural context. Yet to date, the role of internal cultural factors in shaping the translation of AIC into strategic change remains largely unexamined in the AI literature, representing a critical missing piece in the nomological network of AIC-enabled value creation.
This study adopts service-dominant logic (SDL) as its theoretical framework, primarily because it posits that value is co-created through operant resources [13]. Based on this, AIC can be regarded as a higher-order operant resource, providing a theoretical foundation for analyzing the relationship. Moreover, it also emphasizes that this process relies on resource integration and institutional transformation [14], thereby offering a process-based explanation. The strategic challenge lies not simply in sensing, but in converting these insights into coordinated, cross-functional actions. Specifically, operant resources create value when integrated with other resources through organizational routines [13]. In AIC-driven value creation, customer responsiveness (CR) represents a routine that operationalizes the sense–interpret–act cycle by converting AIC into coordinated market-oriented actions. Defined as a firm’s ability to sense, interpret, and act on evolving customer needs [15], it aligns with SDL’s tenet.
A further premise of SDL is that value creation is enabled and constrained by institutional arrangements, or shared norms that coordinate resource integration. In the digital context, we identify digital organizational culture (DOC) as a manifestation of such arrangements, defined as shared assumptions about organizational functioning in digital contexts [16,17]. Existing research has focused on external conditions such as environmental dynamism [18], or competitive intensity [19] as triggers for transformation, elucidating when and why firms need to respond. This study shifts the analytical focus to institutional arrangements, proposing that DOC influences the execution of strategic intent, shaping how AIC is interpreted, shared, and translated into action.
Therefore, this study investigates two research questions: (1) from an SDL value co-creation perspective, whether and how AI capability is linked to business model innovation through the market-oriented mechanism of customer responsiveness; (2) drawing on SDL’s institutional axiom, whether digital organizational culture strengthens the AIC → CR pathway and, in turn, the indirect association between AIC and BMI. By empirically testing these questions, this study offers several theoretical contributions. First, it identifies CR as a mechanism that helps explain the association between AIC and BMI by connecting firm–customer interactions to business model reconfiguration, complementing existing research focused on operational efficiency and resource optimization. Second, it extends SDL to the digital context by specifying how shared norms and values influence the conversion of probabilistic AI outputs into strategic responses. Third, it enriches the nomological network of AIC by introducing DOC as a boundary condition specific to AI contexts, responding to calls for attention to the internal interpretive mechanisms that shape AIC-enabled value creation.

2. Theoretical Background and Literature Review

2.1. Service-Dominant Logic

SDL is a theoretical framework that emphasizes value co-creation via resource integration in dynamic systems [14,20]. It categorizes resources into operand and operant types. Unlike traditional goods-dominant logic which focuses on tangible product exchanges, SDL posits that value stems from operant rather than operand resource applications.
SDL’s core principles are consolidated into five axioms. A key premise holds that operant resources are the primary source of value creation [13], while the fifth axiom emphasizes that value co-creation is enabled and constrained by institutional arrangements [14], defined as the shared norms, rules, and practices that coordinate resource integration. Prior research has applied these ideas to digital contexts in several ways. For instance, Vargo et al. (2024) [21] draw on this premise to examine how digital technologies, through resource liquification, function as operant resources that enable novel forms of value co-creation in service ecosystems. Linde et al. [4] invoked the fifth axiom to investigate how organizational culture shapes digital transformation outcomes. These studies have advanced our understanding of both operant resources and institutional arrangements in digital contexts.
SDL argues that value co-creation follows a repeated sense–interpret–act cycle based on resource integration. As a higher-order operant resource, AIC supplies massive probabilistic market data (sensing foundation); customer responsiveness (CR) acts as an organizational routine to interpret AI data and coordinate cross-functional market actions; digital organizational culture (DOC), as an institutional norm, determines whether firms dare to act on uncertain AI insights. This three-stage cycle perfectly matches our research framework and provides a holistic SDL theoretical lens for AI-driven business model transformation.
Accordingly, we ground our framework in two SDL axioms. Based on this, AIC acts as a higher-order operator resource [21]. Following the logic that value emerges from resource integration, AIC’s potential is actualized through the sense–interpret–act cycle [22]. These provide a theoretical lens to explore how resources facilitate strategic value reconfiguration through organizational translation, and relevant institutional conditions. Traditional digital technologies (e.g., ERP, big data platforms) generate deterministic, clear operational data with definite interpretation standards. In contrast, AI outputs are probabilistic, predictive, and ambiguous, requiring organizational consensus, cross-departmental data sharing and experimental tolerance to realize value. This unique feature makes internal institutional culture (DOC) an indispensable boundary condition ignored in prior transformation research based on SDL.
It is important to clarify the theoretical role that SDL plays in this study. SDL provides a macro-theoretical lens that sensitizes us to two key premises: (a) value emerges from the integration of operant resources, and (b) such integration is shaped by institutional arrangements. However, testing the full scope of SDL, particularly its recursive, feedback-driven dynamics, would require longitudinal or simulation-based designs that track resource integration over extended time horizons. The present study does not attempt to empirically model the entire sense–interpret–act cycle as a dynamic process. Rather, we translate SDL’s core premises into a testable, cross-sectional mediation–moderation framework that isolates one dominant pathway: from AIC (operant resource) through CR (market-oriented integration routine) to BMI (value system reconfiguration). This approach allows us to empirically examine the existence and strength of this pathway, while acknowledging that the broader service-ecosystem dynamics remain a rich avenue for future research. In doing so, we treat SDL as a theoretical foundation for construct selection and hypothesis derivation, rather than as a blueprint for dynamic process modeling.

2.2. AI Capability

Organizational capabilities are routines that combine skills, knowledge, and coordinated actions to generate competitive advantages [23,24]. Applied to AI, AIC represents the capacity to strategically integrate AI infrastructure, data assets, human expertise, and adaptive processes for operational transformation and sustainable advantage. Prior studies have identified various strategic and technical facets of AIC, providing a baseline for understanding its role in modern organizations [25].
Drawing on established frameworks, AIC is conceptualized as a multidimensional construct comprising three integrated facets. Infrastructure capabilities encompass technical resources, such as cloud platforms and algorithms that relates to real-time data processing, critical for high-volume transactions [26]. Business-spanning capabilities represent intangible resources such as AI-driven processes that translate data insights into new value propositions [27]. A proactive stance reflects human capital’s adaptability and data literacy, which are necessary for fostering AI experimentation and cross-functional collaboration [28]. These dimensions collectively capture the infrastructural, processual, and behavioral facets of AIC. Specifically, infrastructure capabilities (ICs) provide basic data operand resources to support market sensing. Business-spanning capabilities (BSs) integrate AI technical resources with business value logic, realizing primary resource integration. Proactive stance (PS) reflects organizational operant resource willingness to continuously experiment and integrate AI resources with customer market resources.

2.3. The Impact of AI Capability on Business Model Innovation

BMI refers to the fundamental reconfiguration of how firms create, deliver, and capture value [29]. Amid intensifying digital competition, established business models can no longer align with the external environment, failing to sustain value creation and instead giving rise to internal inertia and risk. Consequently, BMI has emerged as a critical pathway for overcoming stagnation and addressing industry disruptions [1].
Theoretically, scholars have developed multiple frameworks to explain BMI. Research based on dynamic capability theory suggests that BMI serves as a strategic response to environmental change [30]. The resource-based view emphasizes that unique resource combinations are key drivers of BMI [31]. Meanwhile, the SDL perspective posits that BMI represents a reconfiguration of value co-creation systems [14]. These theoretical lenses provide a multidimensional framework for understanding BMI. In empirical research, existing findings have mainly explored three categories of antecedents: environmental [3], organizational [32], and technological [33] factors.
Consistent with the SDL premise, AIC enables firms to reconfigure their value creation, delivery, and capture systems. From a resource-based view, AIC provides the material foundation for BMI by combining infrastructure, data assets, and human expertise into unique organizational resources. From a dynamic capabilities perspective, AIC allows firms to rapidly adapt to shifting market demands through mechanisms such as algorithm iteration and real-time data analysis [34]. Together, these perspectives suggest that AIC is positively associated with BMI by enabling firms to transform how value is created, delivered, and captured. Accordingly, we propose:
H1. 
AIC has a positive association with BMI.

2.4. The Mediating Role of Customer Responsiveness

CR originates from the Day [35] framework of market sensing and customer-linking capabilities, defined as the ability to sense, interpret, and act on evolving customer needs. It enables organizations to convert market intelligence into coordinated action. Distinct from organizational agility (emphasizing response speed) and absorptive capacity (focusing on knowledge assimilation), CR captures the market-driven process from insight to action, serving as a dynamic routine reflecting the firm’s external market orientation.
It is important to distinguish CR from the outcomes it enables. While CR captures the organizational process of sensing and acting on customer needs (e.g., CR4: “expand into new regional or international markets”), BMI captures the structural reconfiguration of the business model (e.g., VOA4: “a new sales market is established”). In other words, CR4 captures the firm’s responsiveness capability to pursue geographic expansion, whereas VOA4 captures whether such expansion has actually resulted in a structural change to the business model’s value-offering architecture.
From a capability transformation perspective, AIC enhances CR by reinforcing its mechanisms. The three dimensions of AIC contribute to this process. Infrastructure capabilities enable real-time data collection and pattern recognition, providing the technological foundation for market sensing [26]. Business-spanning capabilities transform data into business insights, with machine learning identifying consumption pattern shifts to deepen demand interpretation [27]. A proactive stance ensures agility in executing responses, as AI-literate employees adjust strategies and cross-functional collaboration facilitates insight dissemination [36,37]. These dimensions form a dynamically reinforcing loop: infrastructure enables sensing, business spanning supports deep interpretation, and proactive stance ensures swift action, with new data generated from actions further enhancing sensing capabilities [38]. Based on this analysis, we propose:
H2. 
AIC has a positive association with CR.
From the value reconstruction perspective, CR serves as the mechanism through which AIC manifests in BMI. As a dynamic capability, CR transforms value-creation architecture across three core dimensions that collectively define BMI. First, CR drives value proposition innovation by translating customer insights into new offerings. Firms leverage real-time feedback to develop solutions for evolving needs [39]. Second, CR facilitates value delivery reconfiguration by converting customer data into optimized touchpoints and seamless omnichannel experiences [40]. In addition, CR evolves value capture mechanisms by identifying willingness-to-pay patterns for innovative pricing models [41]. These architectural changes, spanning value propositions, delivery, and capture, collectively constitute BMI. From SDL’s value co-creation system reconfiguration logic, BMI essentially represents the reconstruction of firm–customer resource exchange relationships. AIC alone cannot directly reconstruct value exchange; it must rely on CR to realize bidirectional resource integration between enterprises and customers: firms use AI data resources to identify customer demand resources, then deliver customized value propositions, distribution channels and pricing mechanisms, ultimately forming new business model architecture. Therefore, CR acts as a mandatory intermediate routine linking AIC operant resources to systemic BMI reconfiguration, forming partial mediation. Thus, CR is the key mechanism linking AIC to strategic value reconfiguration. Based on this analysis, we propose:
H3. 
CR has a positive association with BMI.
H4. 
CR mediates the positive relationship between AIC and BMI.
Taken together, the above hypotheses articulate a theoretically grounded nomological network in which CR serves as a key conduit through which AIC is linked to BMI. It is worth emphasizing that these hypotheses, as formulated, are statements about statistical associations and indirect effects in a theoretically specified direction. While the underlying theory (SDL) implies causal processes, the empirical design of this study—based on temporally separated but still perceptual survey data—tests for patterned covariation consistent with these theoretical expectations, rather than establishing definitive causal effects. We therefore interpret our findings as evidence of theory-consistent associations, with causal inferences requiring further validation through alternative research designs.

2.5. The Moderating Role of Digital Organizational Culture

SDL’s fifth axiom emphasizes that value creation is enabled and constrained by institutional arrangements. We argue that DOC serves precisely this institutional role in the AI context, shaping how the probabilistic, ambiguous insights generated by AIC are interpreted, shared, and translated into customer-responsive actions. DOC, as the cultural infrastructure comprising shared norms of data-driven collaboration and experimentation [16,17], determines whether AI insights will remain latent or become active operant resources. Consistent with SDL’s fifth axiom, institutional norms (DOC) constrain the full chain of resource integration: low DOC leads to internal data silos, resistance to AI experimental results and delayed customer response actions, weakening both the AIC → CR path and the subsequent CR → BMI value reconstruction. High DOC forms shared data-driven decision norms, accelerating the transformation of AI probabilistic insights into customer responsiveness, which further amplifies the indirect impact of AIC on BMI. Thus, DOC positively moderates the conditional indirect effect (moderated mediation), supporting H5b. Based on this, we propose:
H5a. 
DOC positively moderates the relationship between AIC and CR, such that the effect is stronger when DOC is high.
H5b. 
DOC positively moderates the indirect effect of AIC on BMI via CR, such that the mediation effect is stronger when DOC is high.

3. Research Design

3.1. Selection of Research Objects

This study targets Chinese firms in digitally intensive sectors, including scientific research and technical services, information transmission, software and IT services, and leasing and business services. These sectors were selected for two reasons: high digital penetration and technological innovation, with notable AI technology application and BMI, ensuring research representativeness. Additionally, their business models inherently rely on data analysis and user behavior insights, making them ideal for the study. The research model is presented in search Model (Figure 1).

3.2. Data Collection

This study employed a three-wave longitudinal design with two-week intervals. We initially distributed the questionnaire to 400 managers and technical staff at Chinese data-intensive enterprises (Time 1), measuring demographics, AIC and DOC, and received 284 valid responses (71% response rate). Two weeks later (Time 2), we surveyed the same respondents to measure CR, and received 229 valid responses. After another two weeks (Time 3), we measured BMI and received 193 valid responses. The final effective sample size is 193. This multi-wave design is intended to reduce common method bias and allow a better temporal ordering of constructs, although we acknowledge that the time interval is relatively short for fully establishing causality. To alleviate causal ambiguity caused by the short two-week intervals, three waves strictly separate measurement timing: Time 1 measures the independent variable (AIC) and moderator (DOC); Time 2 measures the mediator (CR); and Time 3 measures the dependent variable (BMI). This strict temporal separation of predictor, mediator and outcome complies with longitudinal survey standards for causal inference and weakens reverse causality risks. Selected digital-intensive industries rely heavily on data–user interaction and cross-party resource integration, which aligns perfectly with SDL’s service ecosystem and value co-creation core assumptions, ensuring the research context fits the theoretical framework.

4. Data Analysis and Results

4.1. Common Method Bias and Non-Response Bias

Common method bias (CMB) and non-response bias were tested to ensure data quality. First, for CMB, we conducted Harman’s single-factor test, and the results showed that the unrotated first factor explained 37.29% of the total variance, which is well below the 50% threshold. To further confirm this, we also conducted the common latent factor (CLF) test, and the difference between the standardized regression weights with and without the CLF was less than 0.2 for all indicators, confirming that common method bias is not a serious concern in this study. For non-response bias, we compared the demographic and firm characteristics, as well as the key construct scores, between early and late respondents, as well as between respondents who completed all three waves and those who only completed the first wave. No significant differences were found, ruling out non-response bias. Procedural remedies were also adopted to reduce CMB, three-wave staggered data collection, anonymous filling, random ordering of questionnaire items, and clear statement that there were no right or wrong answers to reduce response bias. Statistical tests including Harman’s single-factor test and the common latent factor (CLF) method jointly confirm no severe common method variance.

4.2. Evaluation of the Measurement Model

Due to different criteria for formative and reflective constructs, we employed distinct standards for reliability and validity. With two second-order variables, a two-stage approach was adopted. The repeated indicator method operationalized the higher-order constructs. As both are endogenous, latent variable scores were indicators in the second stage.
To further address potential concerns about conceptual overlap between CR and BMI dimensions, we examined the HTMT values between CR and the four BMI dimensions. The HTMT values ranged from 0.612 to 0.734, all below the conservative threshold of 0.85 [44]. Notably, the HTMT between CR4 and VOA4 was 0.697, indicating that despite surface-level content similarity, the two constructs are empirically distinct. This is consistent with our theoretical positioning of CR as a responsiveness capability and BMI as a business model reconfiguration outcome.
For reflective constructs, we tested reliability, convergent validity, and discriminant validity. As shown in Appendix A, all constructs demonstrated factor loadings over 0.75. Table 1 presents reliability and validity indicators, with Cronbach’s alpha and composite reliability above 0.8, and AVE values exceeding 0.64, indicating good internal consistency and convergent validity. To assess discriminant validity, we compared AVE values with inter-construct correlations. Table 1 illustrates the square root of the AVE for each construct exceeded its correlations with others, confirming adequate discriminant validity. This study further adopts the HTMT criterion to test discriminant validity. All HTMT values between reflective constructs are below the strict threshold of 0.85, which further verifies adequate discriminant validity beyond the Fornell–Larcker criterion.
For formative constructs, we evaluated their validity using weights (significance and relevance), VIF (to assess multicollinearity), and outer loadings (to assess absolute contribution), as shown in Table 2. Following established guidelines for formative measurement [45], we do not report R2 values for the individual dimensions of BMI, because R2 is not a meaningful validity criterion for formative constructs—unlike reflective constructs, where R2 indicates convergent validity. Instead, the adequacy of formative measurement is determined by the significance and magnitude of indicator weights, VIF values below the conservative threshold of 3.0, and outer loadings above 0.50 for indicators with non-significant weights.
Table 3 shows the evaluation of second-order constructs. First, the multicollinearity test revealed that VIF for all first-order dimensions was below 3.0, indicating no serious multicollinearity issues. Second, the weights of the three first-order dimensions of AIC (IC, BS, PS) were 0.373, 0.290, and 0.501, respectively, and all were significant (t > 2.893). The weights of the four dimensions of BMI (VOA, IVC, EVC, FA) were 0.425, 0.211, 0.218, and 0.304, respectively. It is worth noting that two of the BMI dimensions (IVC and EVC) have t-values slightly below 1.96. Similar to the indicator level, this does not invalidate the higher-order construct. Their theoretical relevance as distinct components of BMI, combined with their strong absolute contributions (outer loadings: IVC = 0.731, EVC = 0.746), and low multicollinearity (VIF < 3.0) justify their retention. Two formative indicators (VOA1, FA4) also show non-significant outer weights, which is acceptable for formative constructs for two interrelated reasons. First, these indicators capture distinct and irreplaceable theoretical dimensions of BMI—VOA1 addresses new customer group expansion, while FA4 addresses revenue logic restructuring—and deleting them would fundamentally alter the construct’s content domain. Second, their absolute contributions (outer loadings: VOA1 = 0.573, FA4 = 0.542) exceed the recommended 0.50 threshold, confirming their substantive relevance to the BMI construct. Additionally, all VIF values are below 3.0, eliminating multicollinearity as an alternative explanation for the non-significant weights. Thus, all items are retained.
Blindfolding calculation shows Q2 values of all higher-order constructs exceed 0, confirming sufficient predictive relevance of the measurement model.

4.3. Evaluation of the Structural Model

The structural model evaluation relies on path coefficients, significance levels, R-squared values, and the examination of mediating and moderating effects. We used partial least squares structural equation modeling (PLS-SEM) to test the structural model. We initially included firm size, firm age, and industry sector as control variables; however, none of these variables showed significant associations with the dependent variables, and their inclusion did not alter the significance or magnitude of the hypothesized paths. Following standard practice, we report the more parsimonious model without control variables.
The R2 values were 0.538 for CR and 0.501 for BMI, indicating model coherence. Regarding direct effects, Table 4 shows that AIC is significantly associated with BMI (β = 0.373, p < 0.001), supporting H1, and AIC is significantly related to CR (β = 0.541, p < 0.001), supporting H2. CR is significantly associated with BMI (β = 0.415, p < 0.001), supporting H3. For the mediation effect, we used the bootstrap method with 5000 resamples to examine it. Results show that the indirect effect of AIC on BMI through CR was 0.224 (p < 0.001), supporting H4. The VAF value was 37.5%, calculated as 0.224/(0.224 + 0.373), indicating that CR plays a partial mediating role. For moderation, the interaction between DOC and AIC is significantly associated with CR. For moderation, the interaction between DOC and AIC significant affects CR (β = 0.299, p < 0.001), supporting H5a, indicating DOC strengthens the AIC-to-CR link.
To further test DOC’s moderating effect on CR’s mediation between AIC and BMI, we calculated the conditional indirect effects at different levels of DOC using PLS-SEM bootstrapping with 5000 resamples. Following Hayes’ recommended approach, we computed the index of moderated mediation. We also conducted a supplementary analysis using PROCESS Model 7 to cross-validate the PLS-SEM results [46]; the two methods yielded consistent patterns (see Appendix A for PROCESS output). The results in Table 5 show an index of 0.237 (Boot SE = 0.053, 95% Boot CI [0.141, 0.345]). This indicates a statistically significant moderating effect of DOC on the indirect AIC-BMI path through CR. Conditional indirect effects by DOC level were 0.113 at low DOC (mean − 1 SD, 95% Boot CI [0.050, 0.185]), 0.254 at mean DOC (95% Boot CI [0.176, 0.340]), and 0.396 at high DOC (mean + 1 SD, 95% Boot CI [0.273, 0.526]). This confirms that a stronger DOC amplifies the indirect AIC–BMI effect through CR, supporting H5b.
The conditional indirect effect rises from 0.113 (low DOC) to 0.396 (high DOC), representing a 249% increase in the indirect influence of AIC on BMI via CR. This substantial gap confirms that digital organizational culture acts as a critical institutional enabler of SDL-based cross-party value co-creation, unlocking the full strategic potential of AI capability.

5. Discussion

5.1. Main Findings

This study proposes and tests a research model consisting of several hypotheses: AIC is positively associated with BMI (H1), AIC is positively associated with CR (H2), CR is positively associated with BMI (H3), CR mediates the relationship between AIC and BMI (H4), DOC moderates AIC’s effect on CR (H5a), and the indirect effect of AIC on BMI through CR is moderated by DOC (H5b). Prior AIC-BMI research adopts internal efficiency mechanisms (strategic agility, organizational adaptability) rooted in dynamic capability theory, focusing on intra-firm resource optimization. Vargo et al. (2024) [21] preliminarily discussed digital technologies as operant resources in service ecosystems but treated digital tools as deterministic. This study advances their research by distinguishing ambiguous probabilistic AI outputs from traditional digital technologies and embedding DOC as institutional moderators, forming a targeted AI extension of SDL’s fifth institutional axiom. The main findings are presented below.
Our findings confirm a positive association between AIC and BMI (supporting H1), and further suggest that this association operates, at least in part, through the market-oriented mechanism of customer responsiveness. Specifically, the positive AIC-CR and CR-BMI links (supporting H2 and H3) are consistent with the SDL premise that value emerges not from resources in isolation, but from their integration through organizational routines. The partial mediation documented in H4 (indirect effect = 0.224, VAF = 37.5%) indicates that CR represents one, though not the only, plausible pathway through which AI capability relates to business model reconfiguration.
Additionally, this conversion process is related to a firm’s internal institutional arrangements. The positive interaction between DOC and AIC on CR indicates that “the positive interaction between DOC and AIC on CR suggests that the AIC–CR association is more pronounced under higher levels of digital organizational culture. Moreover, the conditional indirect effect (H5b), which increases substantially from low to high DOC, indicates that the mediating pathway via CR is contextually contingent: its strength depends on whether firms have cultivated shared norms of data sharing, collaboration, and experimentation. These findings are consistent with the interpretation that DOC functions as an institutional enabler that amplifies the realized association between AI capability and customer-directed action, although causal attribution remains subject to the research design’s limitations.

5.2. Theoretical Contributions

This study makes theoretical contributions in several aspects. First, this study integrates AI capability (AIC), service-dominant logic (SDL), and business model innovation (BMI) into a unified, empirically validated framework specifically tailored to the digital context. Prior research on AIC and BMI has predominantly focused on internal efficiency-oriented mechanisms such as resource orchestration and operational optimization. In contrast, this study identifies customer responsiveness (CR) as a critical market-oriented mediating mechanism, providing empirical evidence for how AIC is linked to BMI through dynamic firm–customer interactions, thereby shifting the focus from internal operational improvements to market-facing value co-creation processes and customer-centric value reconfiguration. While CR is a well-established concept in marketing, our contribution lies in configuring it as the missing link within the specific AIC-BMI nomological network, shifting the focus from internal operational improvements to market-facing value co-creation processes.
Second, this study extends service-dominant logic to the unique context of AI-enabled value co-creation, by specifying how shared institutional norms shape the translation of probabilistic AI outputs into actionable organizational action. Prior applications of SDL have predominantly focused on traditional service contexts. The probabilistic and ambiguous nature of AI outputs requires distinct organizational interpretation processes, unlike the deterministic inputs of traditional digital tools. By introducing digital organizational culture (DOC) as a critical institutional moderator, this study demonstrates how shared norms and practices shape the strategic efficacy of AIC as an operant resource.
Third, this study offers a novel internal perspective on the boundary conditions of AIC-enabled value reconfiguration, shifting the focus from prior external environmental factors to internal institutional arrangements. Existing research on the boundary conditions of AIC has predominantly focused on external factors such as environmental dynamism and competitive intensity, which explain when firms need to transform. In contrast, this study provides evidence that digital organizational culture moderates the strength of the association between AI capabilities and strategic action, complementing prior work on external moderators by highlighting the role of internal institutional arrangements.

5.3. Practical Implications

This study makes two practical contributions. For managers and organizations seeking to translate their AI investments into strategic business model transformation, this research provides actionable, systemic insights. It moves beyond the conventional, narrow focus on technology adoption alone, by conceptualizing the translation of AIC into BMI as a holistic, multi-stage organizational process. By constructing and validating this integrated model, we provide managers with a clear roadmap to guide their AI transformation journey.
In addition to technological capabilities, this study reveals that AIC-driven business model transformation requires not only technological capabilities but also a supportive digital organizational culture. When firms establish a data-driven, collaborative culture, AIC is more likely to be interpreted and translated into action. This finding offers managers a pathway for enhancing strategic execution through cultural change, including cultivating cross-functional data literacy, building tolerance for iterative AI experimentation, and embedding evidence-based decision-making into organizational routines, to help firms unlock the full strategic potential of their AI investments.
Besides these, there are the following suggestions: For the AI capability technical layer: Build three-dimensional AIC system including data infrastructure, business AI planning and employee data literacy mechanisms. For the customer value co-creation layer: Establish AI-driven customer sensing-response closed-loop workflows to realize real-time demand interpretation and omnichannel value delivery optimization. For the digital culture institutional layer: Cultivate cross-functional data sharing, AI experimental tolerance and evidence-based decision-making norms to eliminate internal inertia restricting AI value transformation.

5.4. Limitations and Future Research Directions

Our study has several limitations that can be addressed in future research.
First, while our three-wave survey design allows us to examine the proposed relationships with temporal ordering, the short time intervals (two weeks between waves) do not provide sufficient evidence to establish strict causal relationships. Future research could employ longer intervals or experimental designs to better capture causality. Our three-wave design establishes temporal precedence but cannot capture the recursive, feedback-driven dynamics inherent in SDL’s service ecosystem, representing a key avenue for future longitudinal or simulation-based studies.
Second, this study focuses on Chinese firms in digitally intensive sectors. While this allows us to conduct a controlled examination of the proposed relationships in a context with high AI penetration, it limits the generalizability of our findings to other industries or cultural contexts. Future studies could replicate and extend our model in other contexts, such as traditional manufacturing sectors undergoing transformation, or cross-cultural settings.
Third, the measurement of BMI as a formative construct relies on a comprehensive set of dimensions. As noted in our validation analysis, some dimensions (e.g., IVC, EVC) exhibited lower statistical significance, likely due to the complexity of the construct. Future research might refine the measurement scale or use alternative operationalizations of BMI to further validate our findings.
Fourth, all data were collected from single informants (managers/technical staff) via a self-reported survey platform. Although we employed procedural and statistical remedies for common method bias, common method variance remains a potential concern. Future studies could obtain data from multiple informants within each firm or use objective performance indicators to corroborate our perceptual measures.
Fifth, this study does not distinguish general AI and generative AI capabilities. Large generative AI produces stronger ambiguous predictive insights, which may strengthen the moderating role of DOC. Subsequent research can subdivide AI capability types for comparative analysis.
Sixth, and more fundamentally, while this study draws on SDL—a framework that emphasizes recursive, feedback-driven value co-creation processes—the empirical model (PLS-SEM) tests a static, unidirectional mediation–moderation structure. This creates a theoretical–empirical gap: the sense–interpret–act logic underpinning our hypotheses implies dynamic, iterative cycles (e.g., customer-responsive actions generating new data that refine AI models, which in turn enhance future responsiveness), yet our three-wave design captures only a single temporal sequence rather than tracking such recursive dynamics. Consequently, our findings should be interpreted as cross-sectional evidence of theory-consistent associations at a given point in time, rather than as a complete empirical operationalization of SDL’s full service-ecosystem dynamics. Future research could employ experience sampling, daily diary, or computational simulation methods to more faithfully model the iterative feedback loops inherent in AI-enabled value co-creation.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/systems14080933/s1.

Author Contributions

Conceptualization, J.X. and Z.Z.; methodology, J.X. and Z.Z.; software, J.X.; validation, J.X. and Z.Z.; formal analysis, J.X. and Z.Z.; investigation, J.X. and Z.Z.; resources, Z.Z.; data curation, J.X. and Z.Z.; writing—original draft preparation, J.X.; writing—review and editing, J.X. and Z.Z.; visualization, J.X. and Z.Z.; supervision, Z.Z.; project administration, Z.Z.; funding acquisition, Z.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Ethical review and approval were waived for this study, due to this study employs a non-interventional, anonymous questionnaire survey targeting managerial and technical staff in digitally intensive firms. The research content focuses entirely on respondents’ perceptions of their firms’ AI capabilities, customer responsiveness, organizational culture, and business model innovation, without involving any issues related to personal health conditions, diseases, biological samples, or medical experiments. According to Article 32 of the Measures for Ethical Review of Life Sciences and Medical Research Involving Humans (2023), issued by the National Health Commission of China (Document No. 4 [2023]), research using anonymized information data that does not cause harm to the human body, does not involve sensitive personal information, or commercial interests may be exempted from ethical review.

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

Data are contained within the article or Supplementary Materials.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A. Survey Items and Constructs

All survey items adopt a 7-point Likert scale, ranging from 1 = strongly disagree to 7 = strongly agree.
ConstructItemDescriptionEstimate
AI Capability Loading
Infrastructure Capabilities (ICs)IC1We have access to very large, unstructured, or fast-moving data for analysis.0.800
IC2We integrate data from multiple internal sources into a data warehouse or mart for easy access.0.755
IC3We integrate external data with internal to facilitate high-value analysis of our business environment.0.872
IC4We have the capacity to share our data across business units and organizational boundaries.0.835
IC5We are able to prepare and cleanse AI data efficiently and assess data for errors.0.772
Business Spanning (BS)BS1We can develop a clear vision regarding how AI contributes to business value.0.863
BS2We can integrate business strategic planning and AI planning.0.832
BS3We enabled functional area and general management’s ability to understand value of AI investments.0.826
BS4We can establish an effective and flexible AI planning process and developing a robust AI plan.0.825
Proactive Stance (PS)PS1We are capable of and continue to experiment with new AI tools and techniques as necessary.0.823
PS2We have a climate that is supportive of trying out new ways of using AI.0.863
PS3We constantly seek new ways to enhance the effectiveness of AI use.0.879
PS4We constantly keep current with new AI innovations.0.852
Customer Responsiveness Loading
CR1We can respond to changes in aggregate consumer demand.0.827
CR2We can customize a product or service to suit an individual customer.0.817
CR3We can introduce new pricing schedules in response to changes in competitors’ prices.0.790
CR4We can rapidly adjust our market approach when customer preferences shift across regions.0.778
Digital Organizational Culture Loading
DOC1The teams collaborate functionally in the initiatives for the innovation and digital transformation.0.872
DOC2There is a clear orientation to digital technology changes inside the company’s culture.0.816
DOC3The culture of digital innovation and change takes part as a natural process within the Company.0.826
DOC4The organization shares with the staff the digital strategy, taking into consideration their suggestions.0.857
Business Model Innvoation By the implementation of the business model… Weight
Value-Offering Architecture (VOA) VOA1 …new customer groups are addressed. −0.072
VOA2 …new products, services or combinations of both are offered. 0.268
VOA3 …new value-elements are offered. 0.291
VOA4 …a new (sales) market is established. 0.661
Internal
Value Creation (IVC) 		IVC1 …new competencies are required. 0.375
IVC2 …new resources are required. 0.371
IVC3 …new internal processes are established. 0.246
IVC4 …new organizational structures are established. 0.229
External
Value Creation (EVC) 		EVC1 …new sales channels are developed. 0.217
EVC2 …new partnerships are established. 0.240
EVC3 …external partners are integrated into the internal value creation activities in new ways. 0.548
EVC4 …new trade channels are established. 0.215
Financial Architecture (FA) FA1 …a new revenue model is developed. 0.369
FA2 …new cost structures are established. 0.271
FA3 …new cost mechanisms are introduced. 0.345
FA4 …a new revenue logic is applied. 0.210

References

  1. Han, W.; Zhou, Y.; Lu, R. Strategic orientation, business model innovation and corporate performance—Evidence from construction industry. Front. Psychol. 2022, 13, 971654. [Google Scholar] [CrossRef] [Scilit]
  2. Futterer, F.; Schmidt, J.; Heidenreich, S. Effectuation or causation as the key to corporate venture success? Investigating effects of entrepreneurial behaviors on business model innovation and venture performance. Long Range Plan. 2018, 51, 64–81. [Google Scholar] [CrossRef] [Scilit]
  3. Foss, N.J.; Saebi, T. Fifteen years of research on business model innovation: How far have we come, and where should we go? J. Manag. 2017, 43, 200–227. [Google Scholar] [CrossRef] [Scilit]
  4. Linde, L.; Sjödin, D.; Parida, V.; Wincent, J. Dynamic capabilities for ecosystem orchestration A capability-based framework for smart city innovation initiatives. Technol. Forecast. Soc. Change 2021, 166, 120614. [Google Scholar] [CrossRef] [Scilit]
  5. Jorzik, P.; Klein, S.P.; Kanbach, D.K.; Kraus, S. AI-driven business model innovation: A systematic review and research agenda. J. Bus. Res. 2024, 182, 114764. [Google Scholar] [CrossRef] [Scilit]
  6. Wang, N.; Pan, S.; Wang, Y. How can artificial intelligence capabilities empower sustainable business model innovation? A dynamic capability perspective. Bus. Process Manag. J. 2025, 31, 3003–3025. [Google Scholar] [CrossRef] [Scilit]
  7. Abou-Foul, M.; Ruiz-Alba, J.L.; López-Tenorio, P.J. The impact of artificial intelligence capabilities on servitization: The moderating role of absorptive capacity-A dynamic capabilities perspective. J. Bus. Res. 2023, 157, 113609. [Google Scholar] [CrossRef] [Scilit]
  8. Mikalef, P.; Gupta, M. Artificial intelligence capability: Conceptualization, measurement calibration, and empirical study on its impact on organizational creativity and firm performance. Inf. Manag. 2021, 58, 103434. [Google Scholar] [CrossRef] [Scilit]
  9. Shao, S.; Shao, Z.; Xiong, Y. The influence of AI capability on enterprise competitive advantage: The mediating effect of business model innovation. J. Enterp. Inf. Manag. 2026, 1–25. [Google Scholar] [CrossRef] [Scilit]
  10. Ghosh, S. Developing artificial intelligence (AI) capabilities for data-driven business model innovation: Roles of organizational adaptability and leadership. J. Eng. Tech. Manag. 2025, 75, 101851. [Google Scholar] [CrossRef] [Scilit]
  11. Sjödin, D.; Parida, V.; Kohtamäki, M. Artificial intelligence enabling circular business model innovation in digital servitization: Conceptualizing dynamic capabilities, AI capacities, business models and effects. Technol. Forecast. Soc. Change 2023, 197, 122903. [Google Scholar] [CrossRef] [Scilit]
  12. Kumar, V.; Kumar, S.; Chatterjee, S.; Mariani, M. Artificial intelligence (AI) capabilities and the R&D performance of organizations: The moderating role of environmental dynamism. IEEE Trans. Eng. Manag. 2024, 71, 11522–11532. [Google Scholar] [CrossRef] [Scilit]
  13. Vargo, S.L.; Lusch, R.F. Service-dominant logic: Continuing the evolution. J. Acad. Mark. Sci. 2008, 36, 1–10. [Google Scholar] [CrossRef] [Scilit]
  14. Vargo, S.L.; Lusch, R.F. Institutions and axioms: An extension and update of service-dominant logic. J. Acad. Mark. Sci. 2016, 44, 5–23. [Google Scholar] [CrossRef] [Scilit]
  15. Jayachandran, S.; Hewett, K.; Kaufman, P. Customer response capability in a sense-and-respond era: The role of customer knowledge process. J. Acad. Mark. Sci. 2004, 32, 219–233. [Google Scholar] [CrossRef] [Scilit]
  16. Martínez-Caro, E.; Cegarra-Navarro, J.G.; Alfonso-Ruiz, F.J. Digital technologies and firm performance: The role of digital organisational culture. Technol. Forecast. Soc. Change 2020, 154, 119962. [Google Scholar] [CrossRef] [Scilit]
  17. Pangarso, A.; Winarno, A.; Aulia, P.; Ritonga, D.A. Exploring the predictor and the consequence of digital organisational culture: A quantitative investigation using sufficient and necessity approach. Lead. Organ. Dev. J. 2022, 43, 370–385. [Google Scholar] [CrossRef] [Scilit]
  18. Singh, K.; Chatterjee, S.; Mariani, M. Applications of generative AI and future organizational performance: The mediating role of explorative and exploitative innovation and the moderating role of ethical dilemmas and environmental dynamism. Technovation 2024, 133, 103021. [Google Scholar] [CrossRef] [Scilit]
  19. Crick, J.M.; Friske, W.; Morgan, T.A. The relationship between coopetition strategies and company performance under different levels of competitive intensity, market dynamism, and technological turbulence. Ind. Mark. Manag. 2024, 118, 56–77. [Google Scholar] [CrossRef] [Scilit]
  20. Grönroos, C.; Gummerus, J. The service revolution and its marketing implications: Service logic vs service-dominant logic. Manag. Serv. Qual. 2014, 24, 206–229. [Google Scholar] [CrossRef] [Scilit]
  21. Vargo, S.L.; Fehrer, J.A.; Wieland, H.; Nariswari, A. The nature and fundamental elements of digital service innovation. J. Serv. Manag. 2024, 35, 227–252. [Google Scholar] [CrossRef] [Scilit]
  22. Weick, K.E. Making Sense of the Organization, Volume 2: The Impermanent Organization; John Wiley & Sons: Hoboken, NJ, USA, 2009. [Google Scholar]
  23. Spanos, Y.E.; Prastacos, G. Understanding organizational capabilities: Towards a conceptual framework. J. Knowl. Manag. 2004, 8, 31–43. [Google Scholar] [CrossRef] [Scilit]
  24. Helfat, C.E.; Kaul, A.; Ketchen, D.J., Jr.; Barney, J.B.; Chatain, O.; Singh, H. Renewing the resource-based view: New contexts, new concepts, and new methods. Strateg. Manag. J. 2023, 44, 1357–1390. [Google Scholar] [CrossRef] [Scilit]
  25. Desouza, K.C.; Dawson, G.S.; Chenok, D. Designing, developing, and deploying artificial intelligence systems: Lessons from and for the public sector. Bus. Horiz. 2020, 63, 205–213. [Google Scholar] [CrossRef] [Scilit]
  26. Qin, J.; van der Rhee, B.; Venkataraman, V.; Ahmadi, T. The impact of IT infrastructure capability on NPD performance: The roles of market knowledge and innovation process formality. J. Bus. Res. 2021, 133, 252–264. [Google Scholar] [CrossRef] [Scilit]
  27. Joshi, A.; Benitez, J.; Huygh, T.; Ruiz, L.; De Haes, S. Impact of IT governance process capability on business performance: Theory and empirical evidence. Decis. Support Syst. 2022, 153, 113668. [Google Scholar] [CrossRef] [Scilit]
  28. Fosso Wamba, S.; Queiroz, M.M.; Pappas, I.O.; Sullivan, Y. Artificial intelligence capability and firm performance: A sustainable development perspective by the mediating role of data-driven culture. Inf. Syst. Front. 2024, 26, 2189–2203. [Google Scholar] [CrossRef] [Scilit]
  29. Schneider, S.; Spieth, P. Business model innovation: Towards an integrated future research agenda. Int. J. Innov. Manag. 2013, 17, 1340001. [Google Scholar] [CrossRef] [Scilit]
  30. Teece, D.J. Profiting from innovation in the digital economy: Enabling technologies, standards, and licensing models in the wireless world. Res. Policy 2018, 47, 1367–1387. [Google Scholar] [CrossRef] [Scilit]
  31. Clauss, T. Measuring business model innovation: Conceptualization, scale development, and proof of performance. R D Manag. 2017, 47, 385–403. [Google Scholar] [CrossRef] [Scilit]
  32. Snihur, Y.; Wiklund, J. Searching for innovation: Product, process, and business model innovations and search behavior in established firms. Long Range Plan. 2019, 52, 305–325. [Google Scholar] [CrossRef] [Scilit]
  33. Nambisan, S.; Wright, M.; Feldman, M. The digital transformation of innovation and entrepreneurship: Progress, challenges and key themes. Res. Policy 2019, 48, 103773. [Google Scholar] [CrossRef] [Scilit]
  34. Brynjolfsson, E.; McElheran, K. The rapid adoption of data-driven decision-making. Am. Econ. Rev. 2016, 106, 133–139. [Google Scholar] [CrossRef] [Scilit]
  35. Day, G.S. The capabilities of market-driven organizations. J. Mark. 1994, 58, 37–52. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  36. An, M.; Lin, J.; Luo, X.R. The impact of human AI skills on organizational innovation: The moderating role of digital organizational culture. J. Bus. Res. 2024, 182, 114786. [Google Scholar] [CrossRef] [Scilit]
  37. Fountaine, T.; McCarthy, B.; Saleh, T. Building the AI-powered organization. Harv. Bus. Rev. 2019, 97, 62–73. [Google Scholar]
  38. Rust, R.T. The future of marketing. Int. J. Res. Mark. 2020, 37, 15–26. [Google Scholar] [CrossRef] [Scilit]
  39. Payne, A.; Frow, P.; Eggert, A. The customer value proposition: Evolution, development, and application in marketing. J. Acad. Mark. Sci. 2017, 45, 467–489. [Google Scholar] [CrossRef] [Scilit]
  40. Verhoef, P.C.; Broekhuizen, T.; Bart, Y.; Bhattacharya, A.; Dong, J.Q.; Fabian, N.; Haenlein, M. Digital transformation: A multidisciplinary reflection and research agenda. J. Bus. Res. 2021, 122, 889–901. [Google Scholar] [CrossRef] [Scilit]
  41. Syam, N.; Sharma, A. Waiting for a sales renaissance in the fourth industrial revolution: Machine learning and artificial intelligence in sales research and practice. Ind. Mark. Manag. 2018, 69, 135–146. [Google Scholar] [CrossRef] [Scilit]
  42. Mikalef, P.; Islam, N.; Parida, V.; Singh, H.; Altwaijry, N. Artificial intelligence (AI) competencies for organizational performance: A B2B marketing capabilities perspective. J. Bus. Res. 2023, 164, 113998. [Google Scholar] [CrossRef] [Scilit]
  43. Tallon, P.P.; Pinsonneault, A. Competing perspectives on the link between strategic information technology alignment and organizational agility: Insights from a mediation model. Mis Q. 2011, 35, 463–486. [Google Scholar] [CrossRef] [Scilit]
  44. Henseler, J.; Ringle, C.M.; Sarstedt, M. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. Acad. Mark. Sci. 2015, 43, 115–135. [Google Scholar] [CrossRef] [Scilit]
  45. Hair, J.F.; Risher, J.J.; Sarstedt, M.; Ringle, C.M. When to use and how to report the results of PLS-SEM. Eur. Bus. Rev. 2019, 31, 2–24. [Google Scholar] [CrossRef] [Scilit]
  46. Hayes, A.F. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach; Guilford Publications: New York, NY, USA, 2017. [Google Scholar]
Systems 14 00933 g001
Figure 1. Conceptual research model. Notes: Dashed circles represent first-order reflective dimensions of two higher-order formative constructs. AI capability (AIC, formative higher-order construct) includes infrastructure capabilities (IC), business spanning (BS), and proactive stance (PS) [42]. Business model innovation (BMI, formative higher-order construct) consists of value-offering architecture (VOA), internal value creation (IVC), external value creation (EVC), financial architecture (FA) [2]. CR = customer responsiveness (reflective) [43]; DOC = digital organizational culture (reflective) [16]. Firm size, firm age and industry sector are included as control variables. H5a = DOC moderates AIC → CR; H5b = DOC moderates the indirect path AIC → CR → BMI (moderated mediation).
Figure 1. Conceptual research model. Notes: Dashed circles represent first-order reflective dimensions of two higher-order formative constructs. AI capability (AIC, formative higher-order construct) includes infrastructure capabilities (IC), business spanning (BS), and proactive stance (PS) [42]. Business model innovation (BMI, formative higher-order construct) consists of value-offering architecture (VOA), internal value creation (IVC), external value creation (EVC), financial architecture (FA) [2]. CR = customer responsiveness (reflective) [43]; DOC = digital organizational culture (reflective) [16]. Firm size, firm age and industry sector are included as control variables. H5a = DOC moderates AIC → CR; H5b = DOC moderates the indirect path AIC → CR → BMI (moderated mediation).
Systems 14 00933 g001
Table 1. Discriminant validity and reliability of reflective constructs.
Table 1. Discriminant validity and reliability of reflective constructs.
ConstructCronbach’s αAVEC.R. Discriminant Validity (Fornell–Larcker)
ICBSPSCRDOC
IC0.8660.6530.9040.808
BS0.8570.7000.9030.6200.837
PS0.8770.7300.9150.6060.5710.855
CR0.8160.6450.8790.5100.5110.5530.803
DOC0.8640.7110.908 0.307 0.365 0.3160.4990.843
Note(s): C.R. = composite reliability, CR = customer responsiveness (variable).
Table 2. Formative construct validation.
Table 2. Formative construct validation.
ConstructItemsWeightt-ValueVIFOuter Loading (Absolute Contribution)
VOAVOA1−0.0720.7252.1550.573
VOA20.268 *3.0211.8190.801
VOA30.291 *2.8461.9210.824
VOA40.661 ***7.6592.7580.889
IVCIVC10.375 ***4.6472.4270.862
IVC20.371 ***4.7631.8800.855
IVC30.246 *2.9831.8190.799
IVC40.229 *2.7131.7530.787
EVCEVC10.217 *2.2541.8030.779
EVC20.240 *2.8691.8140.802
EVC30.548 ***6.5741.6200.902
EVC40.215 *2.2032.1840.776
FAFA10.369 ***3.4012.2910.849
FA20.271 *2.8742.0520.812
FA30.345 ***3.6612.0010.837
FA40.2101.9261.9370.542
Note(s): VOA, IVC, EVC, and FA are the four dimensions of BMI. Since BMI is modeled as a formative higher-order construct, the primary evaluation criteria are the weights (significance and relevance), VIF (multicollinearity), and outer loadings (absolute contribution). Following the recommendations of Hair et al. (2019) [45] for formative constructs, we do not report R2 for individual dimensions, as R2 is not a meaningful validity criterion for formative measurement. Indicators with non-significant weights (e.g., VOA1, FA4) are retained because (a) their outer loadings exceed the recommended 0.50 threshold, confirming their substantive absolute contribution, and (b) they capture theoretically indispensable facets of BMI (VOA1 addresses new customer groups; FA4 addresses revenue logic restructuring). * p < 0.05; *** p < 0.001.
Table 3. Second-order model results.
Table 3. Second-order model results.
Formative ConstructItemOuter Weightt-ValueVIF
AIC IC0.373 ***3.8981.919
BS0.290 **2.8931.801
PS0.501 ***4.7571.751
BMIVOA0.425 ***3.3992.394
IVC0.2111.7922.623
EVC0.2181.6162.457
FA0.304 **2.5872.316
Note(s): *** p < 0.001, ** p < 0.01.
Table 4. Path coefficient analysis.
Table 4. Path coefficient analysis.
HypothesisCoefficientS.D.t-Valuep-ValueLLCIULCIResult
Direct effect:
AIC → CR		0.5410.04412.1490.0000.4530.626Support
AIC → BMI0.3730.0695.3730.0000.2350.508Support
CR → BMI0.4150.0696.0150.0000.2790.551Support
DOC → CR0.2930.0525.6340.0000.1940.396Support
DOC × AIC → CR0.2990.0505.9970.0000.1910.388Support
Indirect effect:
AIC → CR → BMI		0.2240.0445.5170.0000.1490.307Support
Moderating effect:
DOC × AIC → CR		0.2990.0505.9970.0000.1910.388Support
Table 5. Moderated mediation analysis.
Table 5. Moderated mediation analysis.
Level of DOCConditional Indirect Effect (PLS-SEM Bootstrapping) (AIC → CR → BMI)Bootstrap SE 95% Bootstrap CI Lower 95% Bootstrap CI Upper
Low (Mean − 1 SD)0.1130.0340.0500.185
Mean0.2540.0410.1760.340
High (Mean + 1 SD)0.3960.0650.2730.526
Index of Moderated Mediation0.2370.0530.1410.345
Note: This table reports the primary PLS-SEM bootstrapping results (5000 resamples). As a robustness check, we also conducted a supplementary analysis using PROCESS Model 7 [46]. The point estimates of the conditional indirect effects and the index of moderated mediation were identical to the third decimal place across both methods (see Appendix A for PROCESS output). This consistency demonstrates the high robustness of our findings and confirms that the results are not contingent on the specific analytical platform employed.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xin, J.; Zhang, Z. From Technological Enablement to Value Co-Creation: How AI Capability Is Linked to Business Model Innovation in Digital Firms. Systems 2026, 14, 933. https://doi.org/10.3390/systems14080933

AMA Style

Xin J, Zhang Z. From Technological Enablement to Value Co-Creation: How AI Capability Is Linked to Business Model Innovation in Digital Firms. Systems. 2026; 14(8):933. https://doi.org/10.3390/systems14080933

Chicago/Turabian Style

Xin, Jiayi, and Zhen Zhang. 2026. "From Technological Enablement to Value Co-Creation: How AI Capability Is Linked to Business Model Innovation in Digital Firms" Systems 14, no. 8: 933. https://doi.org/10.3390/systems14080933

APA Style

Xin, J., & Zhang, Z. (2026). From Technological Enablement to Value Co-Creation: How AI Capability Is Linked to Business Model Innovation in Digital Firms. Systems, 14(8), 933. https://doi.org/10.3390/systems14080933

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop