Latest / 1 More Paper / Generative AI as a driver of organizational creativity: Evidence from the absorptive capacity perspective — Boulitama et al. (2026)
Transcript
- You’re listening to “Generative AI as a driver of organizational creativity: Evidence from the absorptive capacity perspective,” by O. Boulitama and colleagues. Published in 2026. ORIGINAL ARTICLE Citation: Boulitama, O., Sabiri, B., Mouchtakir, N., Rahli, D., & Sabri, K. (2026). Generative AI as a driver of organizational creativity: Evidence from the absorptive capacity perspective. Policy, 21, 467–526. Equilibrium. Quarterly Journal of Economics and Economic the linked source Contact to corresponding author: Othman Boulitama, the email address Othman Boulitama Brahim Sabiri Nissrine Mouchtakir Driss Rahli Karim Sabri Generative AI as a driver of organizational creativity: Evidence from the absorptive capacity perspective This is an Open Access article distributed under the terms of the Creative Commons Attribu-tion License (the linked source), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract. Research background: Generative AI (GenAI) is increasingly integrated into organizational routines for information search, idea development, problem-solving, and experimentation. Yet creativity gains from GenAI adoption are not automatic. Firms differ in their ability to recognize valuable AI-generated outputs, interpret them critically, combine them with existing knowledge, and transform them into original and useful solutions. This capability corre-sponds to absorptive capacity (AC) and may explain why similar investments in GenAI produce uneven organizational creativity (OC) outcomes. Purpose of the article: This study examines the relationship between GenAI use and OC from the perspective of AC. It investigates whether AC operates as the main mechanism through which AI-mediated knowledge is translated into creative outcomes. The study also tests whether GenAI has a direct association with OC beyond the capability-building pathway, reflecting its potential to accelerate ideation, broaden exploration, and support experimentation. Methods: The empirical analysis is based on survey data collected from 364 Moroccan firms. The proposed relationships are tested using partial least squares structural equation modeling (PLS-SEM). The model estimates the direct relationship between GenAI use and OC, as well as the indirect relationship mediated by AC. Findings & value added: The findings show that GenAI use is positively associated with AC and that stronger AC is associated with higher OC. GenAI also exhibits a direct positive relationship with OC, suggesting that generative technologies may support faster idea generation, wider exploration of alternatives, and more rapid iterative refinement. The study contributes to the literature by developing and testing a capability-based explanation of GenAI-enabled creativity at the firm level. It indicates that the creative value of GenAI depends not only on access to generative tools but also on organizational routines that enable firms to absorb, evaluate, recombine, and exploit AI-mediated knowledge. Introduction. Organizational creativity (OC) has become a strategic imperative as firms confront accelerated digitalization, platform-based competition, and mounting pressure to differentiate through renewed products, services, processes, and business models. Beyond effi-ciency-oriented performance, organizations are increasingly assessed by their capacity to generate ideas that are both novel and useful and to translate them into implementable outcomes. Foundational research defines creativity as the production of original and useful ideas shaped by expertise, creative-thinking skills, intrinsic motivation, and organizational conditions such as leadership, climate, and resources. From this standpoint, creativity is not merely an individual attribute. It is a systemic outcome of the ways in which organizations structure work, cultivate talent, and govern knowledge flows. The rapid diffusion of artificial intelligence (AI), and particularly generative AI (GenAI), is now reshaping these creative systems. Based on large models capable of producing text, code, images, and other forms of content, GenAI can support ideation, concept refinement, prototyping, and experimentation across business functions. Empirical evidence indicates that GenAI tools may enhance productivity and output quality, especially among novice employees, by providing suggestions, alternative solutions, and contextual information when needed. Experimental work further suggests that GenAI can improve perceived creativity and output quality while reducing the time required for ideation (Dai & van Swol, 2026; Doshi & Hauser, 2024; Heigl, 2025). These findings support the view that GenAI may become a consequential driver of OC and, by extension, firm-level competitiveness. The creative value of GenAI, however, should not be treated as automatic or uniformly positive. Recent studies have increasingly emphasized that the creative benefits of GenAI are accompa-nied by non-trivial organizational and cognitive risks, including the ho-mogenization of ideas, excessive dependence on machine-generated outputs, the gradual erosion of human creative capabilities, and unresolved questions related to ethics, authorship, accountability, and intellectual property. Research on creative and knowledge-intensive industries also suggests that GenAI may reconfigure professional roles and value-creation logics by opening new possibilities for collaboration while simultaneously challenging established professional identities. Nevertheless, existing research still offers an incomplete view of how GenAI translates into creativity at the organizational level. Most studies focus on individual users, specific tasks, or sector-level disruption, ne-glecting the internal dynamics of firms post-GenAI adoption. In particular, we still know relatively little about how organizations interpret AI-generated outputs, combine them with existing knowledge, and convert them into creative routines, decisions, products, or processes. This gap matters because firms may access similar GenAI tools but differ considera-bly in their ability to transform them into organizational creativity. In this respect, absorptive capacity (AC) offers a theoretically rigorous lens for addressing the gap. AC was initially defined as the firm’s ability to recognize the value of external knowledge, assimilate it, and apply it to commercial ends before being reconceptualized as a dynamic capability encompassing acquisition, assimilation, transformation, and exploitation. Meta-analytic evidence indicates that AC is a strong predictor of innovation and knowledge transfer, with performance effects frequently operating through innovation-related outcomes. Recent work further links AC to OC, agility, and resilience under dis-ruptive conditions. In the context of GenAI, this perspective suggests that generative models are not merely content production tools but knowledge infrastructures that expand access to ex-ternalized knowledge, whose creative value depends on how effectively organizations interpret, recombine, and exploit AI-mediated outputs. The growing diffusion of GenAI has consequently stimulated increasing scholarly interest in its implications for creativity and innovation across organizational settings. Existing research indicates that GenAI can support idea generation, problem solving, and creative task performance by expanding access to information and facilitating the recombination of knowledge. Related streams of research have also emphasized the strategic value of AI capabilities for organizational learning, innovation, and performance. Despite these advances, the mechanisms through which organizations transform AI-generated insights into meaningful creative outcomes remain insufficiently understood. The available evidence is still largely concentrated at the individual level, while comparatively less attention has been devoted to OC as a collective and firm-level capability. Moreover, although the effective deployment of GenAI depends on firms’ ability to identify, assimilate, evaluate, and apply new knowledge, the role of AC in shaping the relationship between GenAI and OC has re-ceived limited empirical scrutiny. This limitation is particularly evident in emerging economies, where evidence about the organizational implications of GenAI remains scarce. Against this backdrop, the present study examines GenAI as a driver of OC through the lens of AC. Specifically, it investigates how the use of GenAI in core organizational activities is associated with firms’ AC and how this capability, in turn, relates to OC outcomes. Drawing on the AC perspective, we argue that the creative benefits of GenAI are more likely to emerge when organizations establish robust routines for acquiring AI-related knowledge, critically assessing model-generated outputs, integrating these outputs with existing expertise and organizational processes, and translating them into practical solutions that extend beyond isolated experimentation. AC therefore provides an explanatory mechanism through which organizations may convert the informational and generative potential of GenAI into original and useful ideas. Using survey data from 364 Moroccan firms and partial least squares structural equation modeling (PLS-SEM), this study tests both the direct relationship between GenAI use and OC and the indirect relationship operating through AC. By integrating OC theory with a dynamic capabilities view of AC, the study clarifies the conditions under which GenAI is associated with more original and useful ideas at the firm level, extends GenAI research beyond individual-level experimental settings, and enriches AC scholarship by theorizing and testing its relevance for a new class of gen-eral-purpose AI technologies. The value of the study lies in moving the debate on GenAI and creativity beyond a technology-centered interpretation. Theoretically, it advances a capability-based explanation of GenAI-enabled OC by showing that the creative implications of GenAI depend not only on the generative proper-ties of the technology but also on the organizational capacity to absorb, evaluate, recombine, and exploit AI-mediated knowledge. This positioning connects GenAI research with AC theory and helps explain why similar technological investments may produce different creative outcomes across firms. From a managerial perspective, the study indicates that organizations seeking to derive value from GenAI should complement technological investment with learning routines, knowledge-sharing practices, critical evaluation mechanisms, and governance structures that allow AI-generated outputs to be transformed into original and useful ideas. By providing evidence from Moroccan firms operating in an emerging econo-my, the study also broadens the geographical scope of current research and addresses a context that remains underrepresented in the literature. This paper is organized as follows. The next section reviews the relevant literature and develops the hypotheses. The methodology section presents the research design, measures, data collection, and analytical procedures. The results section reports the empirical findings and robustness checks where applicable. The paper then discusses the implications of the findings for theory and practice and concludes by outlining limitations and directions for future research. Literature review and hypothesis development Organizational creativity as a firm-level capability OC refers to the organizational capacity to generate ideas that are not only novel but also useful, meaningful, and potentially convertible into renewed products, services, processes, and business models. Although creativity was traditionally examined through individual-level lenses, with emphasis on cognitive ability, expertise, intrinsic motivation, and personal creative dispositions, contemporary research increasingly situates creativity within the social, structural, and knowledge-based architecture of the firm. This theoretical shift is consequential because it relocates creativity from the domain of individual ingenuity to the organizational level, where creative outcomes are shaped by the con-figuration of routines, interaction patterns, managerial practices, and knowledge-governance mechanisms. From this perspective, ideas acquire organizational significance only when they are recognized, evaluated, refined, legitimized, and translated into collective outcomes through structured social and organizational processes. From this perspective, OC is produced through the interaction between human agency and organizational design. Leadership practices, organizational climate, incentive systems, cross-functional collaboration, and routinized knowledge exchange influ-ence whether employees are able and willing to formulate unconventional ideas, challenge established assumptions, explore alternative solutions, and recombine knowledge across functional and cognitive boundaries. Firms that institutionalize experimentation, tolerate intelligent failure, and reward knowledge sharing are better positioned to convert dispersed creative potential into implementable ideas (Nguyen Kim & Nguyen Thi Hang, 2024[A1.1]). Such conditions are essen-tial because creative outcomes rarely emerge from novelty alone; they require interpretive spaces, evaluative mechanisms, and organizational routines capable of transforming fragmented insights into coherent and actionable possibilities. This firm-level understanding of OC is particularly relevant in environments characterized by technological turbulence, intensified competi-tion, and the growing strategic importance of knowledge recombination. In such contexts, creativity becomes closely tied to the firm's capacity to reconfigure capabilities, mobilize dis-tributed expertise, and transform heterogeneous knowledge into value-creating innovation. It is precisely within this logic that the role of GenAI becomes analytically significant. If OC is understood as an organizational capability rather than as a purely individual phenom-enon, the creative contribution of GenAI cannot be evaluated solely through its ability to generate content, accelerate ideation, or expand the number of possible solutions. This must be analyzed alongside the organizational mechanisms that interpret, evaluate, contextualize, and embed AI-mediated outputs into creative, innovation-driven practices. This perspective forms the conceptual baseline for integrating GenAI, AC, and OC within a capability-based framework. Generative AI and the reconfiguration of creative work GenAI is progressively reshaping the architecture of creative and knowledge-intensive work by introducing a socio-technical layer through which text, code, images, and other forms of symbolic output can be generated, recombined, and refined at considerable scale. Its relevance for OC lies not merely in the automation of discrete cognitive tasks but in its capacity to widen the cognitive search space available to organizational actors, accelerate exploratory reasoning, and support iterative processes of problem formulation and solution development. Accumulated empirical evidence suggests that GenAI may generate measurable performance gains, although such outcomes remain contingent on the organizational and cognitive conditions under which the technology is used. Empirical evidence from a field study on customer support workflows demonstrates that the integration of GenAI assistance not only drives productivity gains, but also accelerates organizational learning, with particularly pronounced benefits among less experienced workers, indicating that generative tools may codify and diffuse effective practices across organizational settings. Experimental research on creative tasks similarly shows that exposure to GenAI-generated ideas can improve perceived creativity and output quality while reducing the time required for ideation, especially among participants with lower initial creative ability. These empirical patterns also reveal a collective-level concern: when users rely on similar generative systems, creative outputs may become more homogeneous, thereby narrowing diversity and reinforcing convergent patterns of thinking. The conceptual literature in information systems and management further suggests that GenAI should be understood less as a passive content production instrument than as a potential partner in creative and analytical work. By enlarging cognitive search spaces, proposing alternative framings, and enabling the recombination of internal and external knowledge at low marginal cost, GenAI may reshape how organizations explore problems, construct possibilities, and develop new ideas. This argument is consistent with emerging work on organizational innovation, which frames GenAI as a technology capable of supporting both automation and augmentation, depending on how firms redesign workflows, allocate tasks between human and artificial agents, and govern the quality of human-AI interaction. In this sense, the technology in isolation does not determine the creative implications of GenAI. They are mediated by organizational choices concerning task design, knowledge governance, human oversight, and the integration of AI-generated outputs into existing routines. The literature also warns against interpreting GenAI's creative potential in an overly deterministic manner. The value of generative technologies may be weakened by excessive dependence on machine-generated outputs, diminished human judgment, and unresolved issues related to responsibil-ity, authorship, and intellectual property. Accordingly, the central research problem is no longer whether GenAI can produce content but whether organizations can transform AI-mediated outputs into durable creative outcomes at the firm level. Individual-level experiments, task-specific assessments, and broad analyses of sectoral disruption predominantly concentrate the available evidence. Consequently, the existing organizational mechanisms that link GenAI with OC lack sufficient theoretical and empirical testing within an integrated firm-level model. This limitation justifies a capability-based explanation that examines how firms interpret, validate, contextualize, recombine, and exploit AI-mediated knowledge in ways that support OC. Absorptive capacity as the mechanism linking GenAI to creativity AC constitutes a particularly appropriate theoretical lens for explaining how organizations convert externally generated knowledge into creative and innovation-oriented outcomes. In its original formulation, AC refers to the firm’s ability to recognize the value of external knowledge, assimilate it, and apply it to commercial purposes, with prior knowledge and learning routines functioning as essential antecedents of this capability. Zahra and George (2002) later refined this view by conceptualizing AC as a dynamic capability composed of acquisition, assimilation, transformation, and exploitation. This distinction is central to the present study because the creative value of GenAI does not arise from access to AI-generated outputs alone. It depends on the organizational processes through which such outputs are identified as valuable, interpreted in relation to existing knowledge, recombined with internal expertise, and translated into actionable ideas. AC should, therefore, be understood not as a generic learning ability but as a complex organizational capability that connects external knowledge exposure with internal knowledge transformation and application (Durman et al., 2025; Human, 2021; Schweisfurth & Raasch, 2018). The acquisition dimension captures the firm’s capacity to identify and access relevant external knowledge, while assimilation refers to the interpretive processes through which this knowledge is analyzed, understood, and shared across organizational members. Transformation reflects the firm’s ability to recombine newly acquired knowledge with existing cognitive schemas, routines, and experiential knowledge, whereas exploitation concerns the conversion of transformed knowledge into operational, strategic, or innovation-related outcomes. This distinction is particularly important in the context of GenAI because AI-generated outputs are often abundant, probabilis-tic, and context-dependent. Their organizational value depends not only on access to such outputs but also on the firm’s ability to evaluate their relevance, identify their limitations, combine them with domain-specific expertise, and embed them into decisions, processes, or creative initiatives. The literature on innovation has well established the relevance of AC. Synthesized evidence indicates that AC is a robust antecedent of knowledge transfer and innovation, with its performance effects often ma-terializing through innovation-related outcomes rather than through direct efficiency gains. This argument becomes particularly salient in the context of GenAI. Generative technologies substantially increase the volume, speed, and variety of available knowledge inputs. However, informational abundance does not automatically produce organizational creativity. Organizations must establish routines to assess, validate, contextualize, and incorporate AI-mediated outputs into their actions. Without such routines, GenAI may generate numerous suggestions without necessarily improving the originality, relevance, or implementability of creative outcomes. In the context of GenAI, the four dimensions of AC acquire a more specific operational meaning. Acquisition refers to the organization’s capacity to identify relevant GenAI use cases, formulate appropriate prompts, access relevant data sources, and mobilize complementary human expertise. Assimilation concerns the capacity to under-stand AI-generated outputs, evaluate their relevance, detect their limitations, and integrate them into shared organizational interpretations (Alshahrani et al., 2025). Transformation involves the recombination of AI-mediated suggestions with firm-specific knowledge, routines, constraints, and strategic priorities, particularly through iterative refinement across functions. Exploitation refers to the incorporation of validated AI-mediated knowledge into products, services, processes, decisions, and organizational routines through governance, accountability, and operational discipline. GenAI, in this sense, should be understood not simply as a digital tool but as a knowledge-generating infrastructure whose value depends on the firm’s capacity to learn from, challenge, and operationalize its outputs. This reasoning supports the expectation that GenAI use may strengthen AC. By reducing the cost of search and synthesis, accelerating exploratory cycles, and enabling effective responses to be codified and disseminated across employees, GenAI can reinforce the organizational processes through which knowledge is acquired, interpreted, transformed, and ap-plied. Such benefits are more likely to emerge when firms integrate GenAI into core activities rather than confining it to isolated experimentation and when technological adoption is accompanied by training, cross-functional learning routines, and responsible governance. Accordingly, GenAI can be expected to enhance AC not because the technology itself possesses absorptive capacity but because it enlarges and accelerates the knowledge processes through which organizations identify, evaluate, recombine, and exploit knowledge for creative purposes. H1: The use of generative artificial intelligence has a positive effect on firms’ absorptive capacity. From absorptive capacity to organizational creativity AC is theoretically and empirically associated with OC because it en-larges the breadth, depth, and diversity of the firm’s knowledge base while strengthening its capacity to recombine distant and heterogeneous knowledge domains. Through processes of environmental scanning, knowledge assimilation, and interpretive transformation, firms are better able to reframe existing problems, detect novel opportunities, and generate ideas that satisfy both novelty and usefulness requirements (Cohen & Levinthal, 1990; Zahra & George, 2002). OC can therefore be understood as a proximal and capa-bility-dependent outcome of AC, particularly in environments marked by rapid technological change, market volatility, and increasing pressure to transform dispersed knowledge into actionable innovation. Empirical research further supports this relationship by linking AC to adaptive organizational responses such as agility and resilience, both of which require the ability to interpret external signals and recombine knowledge under changing conditions. In this respect, AC does not merely facilitate incremental improvement; it also enables more substantial forms of cognitive and organizational recombination that nourish creative output. Meta-analytic evidence reinforces this argument by showing that AC is a robust antecedent of innovation outcomes, including radical innovation, because it supports capability reconfiguration, cross-domain knowledge integration, and the conversion of external knowledge into valuable organizational outcomes. Given the transformative nature of GenAI, the strategic importance of AC is profoundly amplified. GenAI can generate large volumes of plausi-ble, diverse, and rapidly produced content, yet OC requires more than the availability of ideas. It depends on the firm’s ability to select, validate, contextualize, and implement those ideas in ways that are aligned with organizational knowledge, strategic intent, and operational constraints. Firms with stronger AC are therefore better positioned to transform GenAI outputs into creative outcomes by combining AI-generated alternatives with local expertise, accumulated experience, and organizational priorities. Without such absorptive capability, GenAI may increase the volume of generated content without necessarily improving the originality, relevance, or implementability of creative outcomes. This reasoning also helps address an important limitation in the current literature. Existing studies on AI and creativity have often emphasized the generative capacity of the technology itself, while paying less attention to the organizational mechanisms that determine whether AI-mediated outputs become meaningful creative resources. By placing AC at the center of the GenAI-OC relationship, the present study responds to this gap and provides a clearer theoretical bridge between technological augmentation and firm-level creativity. AC explains why the same technological affordances may produce different creative outcomes across firms: organizations differ in their ability to absorb external knowledge, critically evaluate new informational inputs, recombine them with internal expertise, and translate them into workable ideas. Accordingly, AC is not a peripheral condition but a central explanatory capability through which GenAI-generated possibilities may be converted into OC. H2: Absorptive capacity positively affects organizational creativity. Direct effects of GenAI on organizational creativity and the hybrid creativity view Although AC constitutes a central explanatory mechanism in the proposed model, GenAI may also be directly associated with OC. This direct pathway is theoretically plausible because generative technologies can expand ideational fluency, accelerate prototyping, and support rapid experimentation by enabling organizational actors to explore a broader range of alternatives within shorter cognitive and temporal cycles. Experimental evidence further indicates that GenAI assistance can improve perceived creativity and output quality while reducing the time required for ideation. These findings suggest that GenAI may contribute to creative outcomes not only through the gradual development of organizational learning routines but also through its more immediate capacity to enrich the repertoire of ideas, framings, and possible solutions available to individuals and teams. At the organizational level, we can comprehend this direct association by examining how GenAI transforms creative workflows. By broadening cognitive search spaces and enabling faster iteration, GenAI may support creative activities across functions such as product development, marketing, design, and strategic analysis. Its contribution to OC may therefore derive from its ability to stimulate exploratory thinking, reduce cognitive fixation, and facilitate the comparison of alternative courses of action. Nevertheless, this pathway should not be interpreted in a technologically deterministic manner. The literature indicates that GenAI-related gains remain contingent on the quality of human judgment, task design, and organizational governance and may be weakened by risks such as output homogenization, excessive reliance on algorithmic suggestions, and diminished critical evaluation. This ambivalence provides the theoretical basis for a hybrid creativity view. In this perspective, creativity is not located exclusively in human cognition or in the generative capacity of AI systems. Rather, it emerges from the interaction between human expertise, organizational routines, and AI-mediated ideation. GenAI may enrich the creative process by proposing novel combinations, alternative framings, and unex-pected solution pathways; however, these outputs acquire organizational value only when they are interpreted, selected, refined, and integrated by human actors within specific organizational contexts. The creative contribution of GenAI, therefore, depends on how firms structure human-AI collaboration, preserve critical oversight, and maintain the eval-uative role of human judgment. This argument also addresses an important limitation in the current literature. Existing studies have predominantly examined GenAI in individual, task-based, or experimental settings, leaving insufficient evidence on whether its creative affordances translate into organizational-level outcomes. By examining the direct associations between GenAI use and OC, and testing the mediated pathway via AC, the present study distinguishes between two complementary mechanisms. The first is an augmentation pathway, through which GenAI directly supports ideation, experimentation, and creative exploration. The second is a capability-based pathway, through which AC enables firms to absorb, evaluate, and transform AI-mediated knowledge into creative outcomes. This dual logic provides a more comprehensive theoretical explanation of how GenAI may be associated with OC in firm-level settings. H3: The use of generative artificial intelligence has a positive effect on organizational creativity. The logic of the preceding hypotheses suggests that the relationship between GenAI use and OC should not be interpreted only as a direct technological effect. GenAI may enlarge the repertoire of ideas available to firms, accelerate exploratory search, and support experimentation, yet the organizational value of these outputs depends on whether firms possess the knowledge-processing routines required to interpret, evaluate, recombine, and apply them. This argument is consistent with research showing that GenAI can enhance individual creative output while also reducing the diversity of ideas when users rely on similar AI-generated suggestions. It also aligns with PLS-SEM mediation logic, which recommends testing indirect effects when theory suggests that an exoge-nous construct influences an outcome through an intervening organizational mechanism. In the present model, AC provides this intervening mechanism. If GenAI use strengthens firms' ability to acquire, assimilate, transform, and exploit knowledge, and if these knowledge-processing capabilities subse-quently support OC, then AC should transmit part of the relationship between GenAI use and OC. This mediation hypothesis is theoretically important because it shifts the explanation from technology adoption alone to capability development. It suggests that GenAI becomes creatively valuable not simply because it produces content, but because firms are able to absorb AI-mediated knowledge, subject it to critical interpretation, combine it with internal expertise, and translate it into original and useful organizational outcomes. Accordingly, the following hypothesis is formulat-ed: H4: Absorptive Capacity mediates the relationship between Generative AI use and Organizational Creativity. Positioning and contribution relative to prior research The literature on GenAI, knowledge processes, and creativity has advanced rapidly, but its theoretical consolidation remains incomplete. Experimental and task-based studies have indicated that GenAI can enhance idea generation, perceived creativity, productivity, and output quality, while also increasing the risk of output convergence and reducing diversity in creative production. These studies are valuable because they isolate the immediate creative affordances of GenAI under controlled conditions. However, their explanatory scope becomes limited when the focus shifts from individual task performance to OC. Firms do not simply receive AI-generated ideas. They must interpret them, circulate them across organizational units, assess their relevance, integrate them with existing knowledge, and convert them into routines, decisions, products, or processes. The organizational translation of GenAI outputs therefore remains insufficiently explained when creativity is examined mainly through short-term experimental outcomes. In fact, innovation management and information systems research has begun to conceptualize GenAI as a technology capable of reshaping work design, innovation processes, and human-AI collaboration. Mariani and Dwivedi (2024) argue that GenAI provides new opportunities for innovation management by altering ideation, experimentation, and knowledge recombination. Holmström and Carroll (2025) further show that organizations may innovate with GenAI through different configurations of automation and augmentation, depending on how they redesign workflows and govern human-AI interaction. This literature provides a more organizationally sensitive understanding of GenAI than studies focused solely on task performance. Its limitation, however, lies in the fact that it often remains conceptual, programmatic, or oriented toward broad implications of technological adoption. It rarely specifies the internal capability mechanism through which AI-generated outputs become creative organizational outcomes. The present study addresses this limitation by positioning AC as the knowledge-processing capability through which GenAI may be linked to OC. Research on AI capabilities and organizational learning provides a further foundation for this argument. Ramaul et al. (2024) indicate that industrial AI capabilities develop through situated organizational learning processes in which firms align technological experimentation with existing routines and knowledge structures. This perspective is particularly relevant because it shifts attention away from technology as an isolated input and toward the organizational processes that make technological value possible. However, broader AI capability research does not fully isolate the speci-ficity of GenAI as a generative and knowledge-producing technology. Nor does it sufficiently explain how AI-mediated knowledge is absorbed, transformed, and converted into creativity as a firm-level outcome. By integrating GenAI, AC, and OC within a single empirical model, the present study offers a more precise explanation of how generative technologies become creatively consequential inside organizations. The methodological profile of the existing literature also reveals important limitations. Experimental studies provide strong internal validity, but their reliance on individual-level tasks, artificial settings, and short-term outcomes restricts their ability to capture organizational routines, knowledge flows, governance mechanisms, and collective decision-making. Conceptual and agenda-setting contributions have advanced the theoretical debate on GenAI and innovation, yet they require empirical validation at the organizational level. Empirical studies on AI capability development have begun to address organizational learning processes, but they often focus on AI capabilities or innovation performance in broad terms rather than on the specific GenAI-OC relationship. These methodological constraints justify the need for firm-level empirical models that distinguish between direct augmentation effects and capability-mediated effects. The contribution of the present study lies in repositioning the GenAI-creativity debate around OC rather than technological affordance alone. The study distinguishes between an augmentation pathway, through which GenAI may directly support ideation, experimentation, and exploratory search, and a capability-based pathway, through which AC enables firms to interpret, validate, recombine, and exploit AI-mediated knowledge. This distinction explains why firms exposed to similar GenAI affordances may not achieve comparable creative outcomes. The difference lies not only in access to technology but also in the organizational capacity to absorb and transform what the technology produces. By relying on firm-level survey data and PLS-SEM in the Moroccan context, the study also addresses the limited representation of emerging-market firms in existing research and provides empirical evidence from a setting marked by growing digitalization pressures, uneven technological maturity, and heterogeneous organizational learning routines. Methods. Research design and model specification This study adopts a quantitative explanatory design to examine how GenAI use is associated with OC at the firm level and to clarify the organizational mechanism through which this relationship unfolds. The research design is grounded in the premise that the creative implications of GenAI cannot be reduced to the mere availability of generative tools or to their immediate capacity to produce content, suggestions, or alternative solutions. Although GenAI may directly enrich ideation and experimentation, its contribution to firm-level creativity is expected to depend substantially on the organization’s ability to absorb, interpret, recombine, and exploit AI-mediated knowledge. AC is therefore positioned as the central knowledge-processing capability that links the technological affordances of GenAI to creative organizational outcomes. The conceptual model presented in Figure 1 specifies four theoretically derived hypotheses. The first path links GenAI use to AC (H1), reflecting the assumption that generative technologies may strengthen firms’ capacity to access diverse information, synthesize dispersed knowledge, and support sensemaking processes. The second path links AC to OC (H2), based on the argument that firms capable of acquiring, assimilating, transforming, and exploiting knowledge are better equipped to generate ideas that are both novel and useful. The third path specifies a direct association between GenAI use and OC (H3), capturing the possibility that GenAI may contribute to creativity by broadening the ideational search space, accelerating prototyping, stimulating alternative framings, and supporting exploratory problem-solving. The fourth hypothesis explicitly formalizes the mediating role of AC (H4), suggesting that part of the relationship between GenAI use and OC is transmitted through the firm’s capacity to transform AI-mediated outputs into organizationally meaningful knowledge. This specification distinguishes between two complementary explanatory pathways. The direct pathway captures the augmentation logic of GenAI, whereby generative systems may support creative work by expanding the range, speed, and variety of ideas available to organizational actors. The indirect pathway captures the capability logic of the model, whereby GenAI becomes creatively consequential only when firms possess the routines required to evaluate, contextualize, recombine, and apply the knowledge generated or supported by AI systems. This distinction is important because it prevents a technologically deterministic interpretation of GenAI-enabled creativity. It recognizes that the same technological affordances may produce different creative outcomes depending on the organizational capability to convert AI-mediated knowledge into actionable and original ideas. Given the cross-sectional nature of the data, the empirical strategy is designed to test theoretically grounded associations and mediation patterns rather than to establish definitive causal relationships. This warning is especially important in the GenAI context, where reverse causality and unob-served heterogeneity are still possible. Firms with stronger prior creativity, higher digital maturity, or more developed learning routines may be more inclined to adopt GenAI, and they may also be better positioned to develop AC. Accordingly, the study emphasizes theoretical consistency, construct validity, transparent model specification, and cautious interpretation of the estimated relationships. The objective is not to claim causal certainty but to assess whether the observed empirical structure is consistent with the proposed capability-based explanation of GenAI-enabled OC. To test the model, the study applies partial least squares structural equation modeling (PLS-SEM). This approach is appropriate because the model includes latent organizational constructs, theoretically specified direct paths, and an explicit mediation mechanism linking GenAI use, AC, and OC. PLS-SEM enables the simultaneous assessment of the measurement model and the structural model, including reliability, convergent validity, discriminant validity, explanatory power, predictive relevance, direct effects, indirect effects, and effect sizes. The final structural specification focuses exclusively on the hypothesized relationships among GenAI use, AC, and OC, without control variables, in order to preserve the theoretical focus of the model and ensure alignment between the conceptual framework, the empirical estimation, and the interpretation of the findings. Sample and data collection Data were collected through a structured survey administered to Moroccan firms, with the complete questionnaire reported in Appendix 1. The final dataset comprises 364 valid firm-level responses. Morocco provides a relevant empirical context for examining GenAI-enabled creativity because firms operate under increasing digitalization pressure while display-ing heterogeneous levels of technological maturity, governance capacity, and organizational learning routines. This contextual diversity makes it possible to examine how GenAI use, AC, and OC are articulated across firms with different organizational profiles, without restricting the analysis to a single industry or technological domain. The study relied on a non-probability purposive sampling strategy tar-geting Moroccan firms operating across different economic sectors. This approach was appropriate because the research examines organizational practices related to GenAI integration, AC, and OC, which require respondents who are sufficiently familiar with their firm’s digital practices, knowledge-processing routines, and creativity-related activities. Respondents were therefore selected as key informants rather than as ordinary users of digital tools. Recruitment was conducted through professional net-works, direct organizational contacts, and online communication channels. To be retained in the sample, respondents had to belong to an active Moroccan firm and be able to provide informed assessments of GenAI use, absorptive capacity, and organizational creativity at the organizational level. Incomplete questionnaires, inconsistent responses, and low-quality response patterns were excluded from the final dataset. The target sample size was initially assessed using Roscoe (1975) guide-line, according to which the sample should include at least ten observations per measurement item. Since the initial instrument included 15 items, the minimum required sample size was 150 firms. To reinforce the empirical robustness of the study, 400 Moroccan firms were approached, of which 364 provided valid responses, corresponding to a response rate of 91%. This high response rate should be interpreted in light of the targeted re-cruitment process. The sampling frame was not based on a broad anony-mous mailing list but on the purposive identification of firms likely to have sufficient exposure to digital practices, knowledge-management routines, or emerging uses of GenAI. Direct contact, prior professional accessibility, and systematic follow-up helped reduce non-response and explain the high completion rate. A statistical power analysis was also conducted to assess the adequacy of the final sample size for testing the structural model. Following the logic of regression-based power analysis commonly used in PLS-SEM, the max-imum number of predictors directed toward an endogenous construct was considered. In the present model, the highest number of predictors is two, corresponding to the paths from GenAI use and AC to OC. Using GPower parameters for a linear multiple regression model, with a medium effect size of f2 = 0.15, a significance level of α = 0.05, statistical power of 0.80, and two predictors, the minimum required sample size is approximately 68 observations. Since the final dataset includes 364 valid responses, the sample size substantially exceeds this requirement, indicating that the study has sufficient statistical power to detect medium effects in the structural model. To maximize coverage and improve data quality, three complementary modes of data collection were used. PAPI enabled the in-person admin-istration of paper-based questionnaires, facilitating direct interaction with respondents and allowing clarification when necessary. CAWI was imple-mented through an online platform, which made it possible to reach geo-graphically dispersed firms while reducing data-entry errors. CATI was used to follow up with hard-to-reach respondents and improve response completeness and consistency. The same questionnaire, item wording, response scales, and coding procedures were used across all three modes to enhance comparability and reduce measurement differences. Data collection was conducted over four months, from July 2025 to October 2025. Because the study relied on PAPI, CAWI, and CATI, the possibility of mode-related effects was considered. A direct statistical comparison of construct scores across the three modes could not be performed because the final dataset did not retain respondent-level information on the mode through which each questionnaire was collected. This limitation is acknowledged explicitly. As a complementary robustness check, non-response bias was assessed using the early-late respondent approach, in which the first 25% of responses were compared with the last 25% through independent-samples t-tests on the main construct scores. Late respondents are often considered closer to non-respondents than early respondents, making this comparison a useful diagnostic for potential non-response bias. The results showed no statistically significant differences between early and late respondents for the three main constructs. For OC, early respondents reported a mean score of -0.1485, compared with -0.0087 for late respondents (t = -0.8428; p = 0.4005). For GenAI use, the difference was also non-significant, with early respondents reporting a mean of -0.1657 and late respondents a mean of -0.0007 (t = -1.0291; p = 0.3049). Similarly, no significant difference was observed for AC, with early respondents reporting a mean of -0.0755 and late respondents a mean of 0.0121 (t = -0.5644; p = 0.5732). Since all p-values exceed the 0.05 threshold, non-response bias does not appear to represent a serious concern in this study. However, response timing should not be interpreted as a direct substitute for a collec-tion-mode comparison, and the absence of respondent-level mode information remains a methodological limitation. Respondents were managers and decision-makers with substantive knowledge of GenAI use, organizational knowledge-management practices, and creativity- and innovation-related activities within their firms. Given the firm-level nature of the constructs, the key informant approach was appropriate, although it may introduce a risk of single-source bias. To reduce this risk, the questionnaire items were designed, where possible, to capture observable organizational practices and routines rather than purely subjective evaluations. In addition, several procedural remedies were im-plemented to reduce the risk of common method bias during data collection. Respondents were assured that their answers would remain anony-mous and confidential; the items measuring the independent, mediating, and dependent constructs were separated in the questionnaire, and the order of questions was arranged to reduce consistency-seeking behavior, repetitive response patterns, and automatic answering. Clear instructions emphasized that there were no right or wrong answers and that respondents should answer according to their actual organizational perceptions and practices. All data were used exclusively for academic research purposes, analyzed in aggregated form, and reported in a way that prevented the identification of any individual respondent or firm. Before model estimation, the dataset was screened for completeness and response quality. Observations with substantial missing data or response patterns indicating low quality, such as uniform answers across items, were excluded. The final dataset of 364 valid observations was retained for hypothesis testing and analyzed using ADANCO, which is suitable for PLS-based structural equation modeling and consistent with established methodological standards. Measures and operationalization All focal constructs were operationalized as latent variables measured through multi-item Likert scales, as reported in Table 1. The measurement model was specified reflectively because GenAI use, AC, and OC are con-ceptualized as latent organizational phenomena manifested through ob-servable indicators. The indicators were therefore treated as empirical ex-pressions of broader organizational attributes rather than as independent components forming composite indexes. This specification is consistent with reflective measurement logic, according to which variation in the latent construct is expected to be reflected across conceptually related indicators that share a common domain and display internal consistency. A formative or composite specification would have been more appropriate if the indicators represented distinct, non-interchangeable components whose combination constituted the construct. This was not the logic adopt-ed here. AC was treated as an underlying capability reflected through knowledge acquisition, assimilation, transformation, and exploitation routines, while OC was treated as a latent organizational condition reflected through idea generation, experimentation, and creative problem-solving practices. The measurement items were adapted from established literature and contextualized to the GenAI setting while preserving their conceptual meaning. GenAI use captures the extent to which generative technologies are embedded in organizational activities such as information search, content generation, ideation support, prototyping, and iterative refinement. In this study, GenAI use was not operationalized as an individual-level ac-ceptance construct but as a perceptual measure of organizational integration. Although the questionnaire was completed by individual respondents, the items were formulated to capture how GenAI is incorporated into organizational practices, work routines, decision-making processes, and creativity-related activities. Respondents were therefore treated as key informants reporting on firm-level use rather than on their personal ac-ceptance of the technology. This adaptation shifts the focus from individual attitudes toward GenAI to the degree of its integration into organizational activities and knowledge processes. The constructs were specified as unidimensional reflective measures because the study examines their overall theoretical role within the structural model rather than the separate contribution of each internal subdimension. This choice is particularly relevant for AC. Although AC is theoretically composed of acquisition, assimilation, transformation, and exploitation, the present study conceptualizes it as an overarching organizational capability reflected through these interrelated knowledge routines. A multidimen-sional or higher-order specification would be more appropriate if the objective were to compare the relative contribution of each AC dimension. Since the present model tests AC as an integrated mediating capability between GenAI use and OC, a unidimensional reflective specification was retained for reasons of theoretical parsimony, model clarity, and consistency with the hypothesized mediation logic. AC reflects the organization’s capability to acquire, assimilate, transform, and exploit knowledge. Its operationalization follows the dynamic capability view of AC, which emphasizes that knowledge creates value only when firms move beyond access to information toward interpretation, recombination, and application. In the GenAI context, acquisition refers to the identification of relevant use cases and valuable externalized knowledge; assimilation concerns the interpretation and evaluation of model-generated outputs; transformation captures the recombination of AI-mediated suggestions with firm-specific expertise, routines, and constraints; and exploitation reflects the embedding of validated outputs into processes, services, products, or decision routines. The instrument therefore focuses on learning routines and organizational behaviors rather than on the mere availability of GenAI tools. OC denotes a firm’s capacity to generate ideas that are both novel and useful and to translate them into operational outcomes. To preserve conceptual separation from downstream commercialization, the operationalization focuses on creative outputs and internal creative processes, including problem reframing, option generation, experimentation, and creative problem solving, rather than on market-launch indicators alone. To avoid construct contamination between GenAI use and OC, the OC scale was revised by removing the item that explicitly referred to GenAI. The retained items therefore measure OC independently from any specific technological tool, strengthening the validity of the dependent variable. Firm size, firm age, and industry were collected for descriptive and sample-characterization purposes only. They were not retained as control variables in the final PLS-SEM specification because the empirical model was designed to test the hypothesized direct and indirect relationships among GenAI use, AC, and OC. This decision preserves the theoretical focus of the study and ensures consistency between the conceptual model, the estimated model, and the reported results. The selection of measurement items was guided by the need to balance conceptual coverage, psychometric quality, and survey parsimony. Scale development research recommends using a sufficient number of indicators to represent the construct domain while keeping the instrument concise enough to reduce respondent fatigue and measurement error. Reflective latent constructs are generally well cap-tured by approximately four to six indicators, which supports stable pa-rameter estimation, internal consistency, and convergent validity without introducing unnecessary redundancy. Accordingly, the final measurement model retained five items for GenAI use, five items for AC, and four items for OC, providing a balanced specification between conceptual representativeness and measurement efficiency. Analytical approach: PLS-SEM estimation and hypothesis testing This study adopts a quantitative explanatory design to examine how GenAI use is associated with OC at the firm level and to clarify the organizational mechanism through which this relationship operates. The research design is grounded in the assumption that the creative implications of GenAI cannot be reduced to the mere adoption of generative tools or to their technical capacity to produce content, suggestions, or alternative solutions. While GenAI may directly enrich ideation, experimentation, and exploratory problem-solving, its organizational value is expected to depend on the firm’s capacity to absorb, interpret, recombine, and exploit AI-mediated knowledge. Therefore, AC serves as the primary knowledge-processing capability that links GenAI usage to creative organizational outcomes. As presented in Figure 1, the conceptual model specifies four theoretically derived hypotheses. H1 proposes a positive association between GenAI use and AC, reflecting the idea that generative technologies may support information search, knowledge access, sensemaking, and the synthesis of dispersed insights. H2 proposes a positive association between AC and OC, since firms that are more capable of acquiring, assimilating, transforming, and exploiting knowledge are better positioned to generate ideas that are both novel and useful. H3 specifies a direct association between GenAI use and OC, capturing the possibility that GenAI contributes to creativity by widening the ideational search space, accelerating prototyping, stimulating alternative framings, and supporting exploratory work. H4 formalizes the mediation logic of the model by proposing that AC mediates the relationship between GenAI use and OC. This model distinguishes between two complementary explanatory pathways. The first is a direct augmentation pathway, through which GenAI use may support OC by increasing the range, speed, and diversity of ideas available to organizational actors. The second is an indirect capa-bility-based pathway, through which GenAI use is associated with OC via AC. This distinction is theoretically important because it avoids a deterministic interpretation of GenAI-enabled creativity. It recognizes that access to generative technologies does not automatically produce creative organizational outcomes. Rather, firms must possess the routines required to evaluate AI-generated outputs, contextualize them, recombine them with internal expertise, and convert them into actionable and original ideas. Given the cross-sectional nature of the data, the empirical strategy is designed to test theoretically grounded associations and mediation patterns rather than to establish definitive causal effects. This caution is especially important in the GenAI context, where reverse causality and unobserved heterogeneity are still possible. Firms with stronger prior creativity, higher digital maturity, or more developed learning routines may be more in-clined to adopt GenAI, and they may also be better positioned to develop AC. Accordingly, the study emphasizes theoretical coherence, construct validity, transparent model specification, and cautious interpretation of the estimated relationships. The objective is not to claim causal certainty but to assess whether the observed empirical structure is consistent with the proposed capability-based explanation of GenAI-enabled OC. To test the model, the study applies partial least squares structural equation modeling (PLS-SEM). This approach is appropriate because the model includes latent organizational constructs, theoretically specified direct relationships, and an explicit mediation mechanism linking GenAI use, AC, and OC. PLS-SEM allows the simultaneous assessment of the measurement model and the structural model, including reliability, convergent validity, discriminant validity, explanatory power, predictive relevance, direct effects, indirect effects, and effect sizes. The final structural specification focuses exclusively on the hypothesized relationships among GenAI use, AC, and OC, without control variables, in order to preserve alignment between the conceptual framework, the empirical estimation, and the interpretation of the findings. Methodological strengths and limitations The methodological approach adopted in this study presents several strengths. It examines a theoretically grounded mechanism at the organizational level by linking GenAI use, AC, and OC within a coherent capability-based model. The operationalization of the focal constructs as latent variables helps reduce measurement error and allows the study to assess the reliability and validity of the measurement model before interpreting the structural relationships. The use of PLS-SEM is also appropriate for the proposed framework, given the presence of latent constructs, direct relationships, and an explicit mediation mechanism. In addition, the final sample of 364 firms provides an adequate empirical basis for model estimation, bootstrap-based inference, and the assessment of direct and indirect effects. The study also incorporates several methodological safeguards that strengthen the credibility of the empirical analysis. The questionnaire was administered to key informants with knowledge of their firms’ GenAI use, knowledge-management practices, and creativity-related activities. Procedural remedies were implemented to reduce common method bias, including anonymity, confidentiality, separation of measurement items, careful ordering of questions, and aggregated reporting of results. A statistical common method bias assessment was also conducted using the full collinearity VIF procedure, and the results remained below the recommended threshold. In addition, non-response bias was examined through the early-late respondent approach, with no statistically significant differences observed across the main constructs. These procedures do not eliminate all methodological concerns, but they provide additional reassurance regard-ing the internal consistency and robustness of the survey-based evidence. Several limitations should nevertheless be considered when interpreting the findings. The cross-sectional design restricts the possibility of drawing strong causal conclusions and leaves room for reverse causality and unob-served heterogeneity. Firms with stronger creative capabilities, higher digital maturity, or more advanced knowledge-processing routines may be more likely to adopt GenAI rather than use it alone to explain differences in AC and OC. For this reason, the results should be interpreted as theoretically grounded associations rather than definitive causal effects. Future research should use longitudinal designs to examine how GenAI adoption, AC development, and OC evolve. The use of self-reported perceptual data also represents a methodological limitation, even though it is common in firm-level organizational research. Respondents may overestimate the maturity of GenAI integration, the sophistication of knowledge routines, or the creative capacity of their organizations, particularly because GenAI remains an emerging and highly visible technology. Multi-informant designs, archival indicators, behavioral measures, or externally assessed creative outputs would provide stronger evidence and reduce reliance on single-source perceptions. Moreover, although the study used PAPI, CAWI, and CATI to improve coverage and data quality, the final dataset did not retain respondent-level information on the mode of collection. Consequently, a direct statistical comparison of mode effects could not be performed, and this point should be acknowledged as a limitation. Finally, the sampling strategy, while appropriate for reaching firms with sufficient exposure to digital practices and GenAI-related activities, may underrepresent organizations with low digital maturity or limited technological readiness. These factors may affect the generalizability of the findings and potentially lead to more favorable estimates of GenAI-related organizational capabilities. The Moroccan setting also provides a valuable emerging-economy context, but the institutional, technological, and managerial conditions observed in this study may not fully generalize to other countries or sectors. Future research could replicate the model across different institutional environments, compare sectors with varying levels of AI maturity, and model AC as a higher-order or multidimensional construct to examine whether acquisition, assimilation, transformation, and exploitation play distinct roles in the GenAI-OC relationship. Results. Sample presentation Although service-oriented and knowledge-intensive activities dominate, the sample exhibits a diversified sectoral profile. Firms operating in the tertiary sector constitute the largest share of the sample (40%), followed by firms from the quaternary sector (34%) and the secondary sector (24%), while primary-sector firms account for a smaller proportion (3%). This distribution is consistent with the focus of the study, since GenAI use, knowledge recombination, and creativity-related routines are more likely to be visible in organizational contexts where information processing and service innovation play a central role. At the same time, the inclusion of firms from several sectors provides a sufficiently heterogeneous empirical basis for examining GenAI use, AC, and OC beyond a single industry setting. The sector variable was used only for descriptive and sample-characterization purposes and was not included as a control variable in the final PLS-SEM specification. In terms of annual revenue, the sample is mainly composed of medium-sized and economically established firms. The largest group of respondents reported annual revenues between 50 and 100 MMAD (43%), followed by firms with revenues between 10 and 50 MMAD (32%). Smaller proportions reported revenues between 150 and 200 MMAD (12%) and between 100 and 150 MMAD (7%), while firms below 10 MMAD and above 200 MMAD represented 3% and 2% of the sample, respectively. This revenue structure indicates that the sample includes firms with sufficient organizational and financial capacity to engage with digital technologies while avoiding excessive concentration among very large firms. The employment-size distribution follows a similar pattern. Firms employing between 51 and 100 employees represent 40% of the sample, followed by firms with 101 to 150 employees (35%). Firms with 10 to 50 employees account for 16%, whereas firms with fewer than 10 employees represent 5%. Larger firms remain less frequent, with 3% employing between 151 and 200 employees and 1% employing more than 200 employees. This composition suggests that the empirical analysis is primarily grounded in firms with enough organizational structure to develop knowledge routines, coordinate digital practices, and support creative activities at the firm level. Firm age presents a relatively balanced distribution. Firms operating for 5 to 10 years account for 28% of the sample, followed by those operating for 11 to 15 years (26%) and 16 to 20 years (24%). Firms with fewer than 5 years of activity represent 21%, while firms operating for more than 20 years account for only 1%. This distribution captures organizations at different stages of maturity, ranging from younger firms still consolidating their routines to more established firms with accumulated organizational experience. Overall, the sample provides a varied empirical basis in terms of sector, revenue, employment size, and organizational maturity, supporting the relevance of examining GenAI use, AC, and OC at the firm level. Assessment of the measurement model Reliability and convergent validity of the measurement model Table 2 reports the results of the measurement model assessment, including indicator loadings, indicator-level VIF values, internal consistency reliability, and convergent validity. The results provide satisfactory evidence of the psychometric quality of the three constructs retained in the revised model. Internal consistency reliability is supported across all constructs. Dijkstra-Henseler’s rho ranges from 0.8413 to 0.8853, Jöreskog’s rho ranges from 0.8400 to 0.8839, and Cronbach’s alpha ranges from 0.8417 to 0.8837. These values exceed the recommended threshold of 0.70, indicating. that the measurement scales display adequate internal consistency and that the retained indicators coherently reflect their respective latent constructs. Convergent validity is also supported by the pattern of indicator loadings and AVE values. For OC, the retained item loadings range from 0.7046 to 0.7844, indicating acceptable indicator reliability after the removal of the item that explicitly referred to GenAI. For GenAI use, loadings range from 0.7343 to 0.8186, while the AC indicators load between 0.7224 and 0.7926. These values indicate that the indicators contribute meaningfully to their respective constructs. The AVE values are also above the recommended threshold of 0.50, with 0.5716 for OC, 0.6040 for GenAI use, and 0.5679 for AC. This confirms that each construct explains more than half of the variance of its indicators on average, thereby supporting convergent validity. Indicator-level multicollinearity was examined using VIF values. The results do not indicate problematic redundancy among the retained indicators. VIF values range from 1.2938 to 3.2057 for OC, from 1.6601 to 2.9898 for GenAI use, and from 1.5854 to 3.4671 for AC. Although one AC indicator approaches the upper range commonly regarded as desirable, the values remain within acceptable limits and do not suggest severe collinearity. The revised measurement model therefore demonstrates adequate reliability, convergent validity, and acceptable indicator-level collinearity, providing a sound basis for the assessment of the structural model. Discriminant validity Table 3 reports discriminant validity using the HTMT criterion, following the recommendations of Henseler et al. (2015). The results provide satisfactory evidence that the three constructs are empirically distinct. The HTMT value between GenAI use and OC is 0.8196, the value between AC and OC is 0.7510, and the value between GenAI use and AC is 0.8000. All values remain below the conservative threshold of 0.85, indicating that the constructs do not exhibit excessive empirical overlap despite their theoretical relatedness. Discriminant validity is therefore established, confirming that GenAI use, AC, and OC capture distinct dimensions of the proposed model and can be interpreted separately in the structural analysis. Structural model assessment The structural model was estimated without control variables, as the final specification was designed to test the hypothesized direct and indirect relationships among GenAI use, AC, and OC. The graphical representation of the estimated model is presented in Figure 2, while the detailed structural results are reported in Table 4. The model explains a substantial pro-portion of variance in the two endogenous constructs. GenAI use accounts for 64.41% of the variance in AC (R2 = 0.6441), while GenAI use and AC jointly explain 69.67% of the variance in OC (R2 = 0.6967). The predictive relevance values are also satisfactory, with Q2 = 0.357 for AC and Q2 = 0.422 for OC, indicating that the model has meaningful predictive relevance for both endogenous constructs. These results suggest that the structural model provides a strong empirical basis for examining the relationships among GenAI use, AC, and OC. The effect-size results reported in Table 4 further clarify the substantive relevance of the structural paths. The effect of GenAI use on AC is large (f2 = 1.8096), confirming that GenAI use is strongly associated with the knowledge-processing routines through which firms acquire, assimilate, transform, and exploit knowledge. The effect of GenAI use on OC is also substantial (f2 = 0.4277), indicating that generative technologies are directly associated with firms’ capacity to generate novel and useful ideas. By con-trast, the effect of AC on OC is positive but more modest (f2 = 0.0846), suggesting that absorptive capacity contributes to organizational creativity, although its incremental effect is weaker than the direct GenAI-OC pathway. The bootstrapping results reported in Table 5 provide further evidence for the hypothesized direct relationships. GenAI use is positively and significantly associated with OC (β = 0.6037, t = 4.1464), supporting H3. GenAI use is also strongly and significantly associated with AC (β = 0.8025, t = 23.1906), supporting H1. The relationship between AC and OC is positive but weaker (β = 0.2685, t = 1.8430), providing weak support for H2. These results indicate that GenAI use is associated with OC through two complementary pathways: a dominant direct augmentation pathway and a weaker capability-based pathway operating through AC. The mediation analysis reported in Table 5 shows that the indirect effect of GenAI use on OC through AC is positive but weakly supported (indirect effect = 0.2155, t = 1.8482). This result provides weak support for H4 and suggests that AC partially mediates the relationship between GenAI use and OC. The total effect of GenAI use on OC, reported in Table 4, is 0.8192, combining a direct effect of 0.6037 and an indirect effect of 0.2155. This pattern suggests that the development of absorptive capacity does not solely explain the creative relevance of GenAI. Rather, GenAI use appears to be associated with OC primarily through its direct contribution to ideation, exploration, and iterative refinement, while AC provides an additional but comparatively weaker mechanism through which AI-mediated knowledge is absorbed, recombined, and translated into creative organizational outcomes. Discussion. GenAI, absorptive capacity, and organizational creativity: Interpretation of the empirical findings The findings indicate that GenAI use is positively associated with OC in Moroccan firms, both directly and indirectly through AC. However, the corrected structural results call for a more nuanced interpretation than a purely mediation-centered explanation. The direct association between GenAI use and OC is the strongest pathway in the model, whereas the indirect pathway through AC is positive but comparatively weaker. This pattern suggests that GenAI is associated with organizational creativity primarily by supporting ideation, exploratory search, alternative framing, and iterative refinement, while AC provides an additional knowledge-processing mechanism through which AI-mediated outputs can be evaluated, recombined, and translated into creative organizational outcomes. Research suggests that integrating GenAI into organizational workflows can reshape ideation, design, and decision-making. The present study expands on this argument by demonstrating that workflow augmentation alone cannot fully capture the creative relevance of GenAI at the firm level. GenAI use is also strongly associated with AC, indicating that firms using generative technologies more systematically tend to report stronger routines for acquiring, assimilating, transforming, and exploiting knowledge. In this respect, GenAI appears to enlarge the informational and cognitive resources available to firms, while AC captures the organizational capacity to interpret and mobilize those resources. The strongest structural association concerns the relationship between GenAI use and AC. This finding indicates that firms integrating GenAI into their core activities are more likely to develop or activate knowledge-processing routines. This result is consistent with Abou-Foul et al. (2023), who argue that AI-enabled systems enhance firms’ ability to capture and process information from their environment, and with Liao et al. (2023), who highlight the role of AI technologies in strengthening knowledge acquisition and utilization. In the GenAI context, this relationship is particularly meaningful because generative tools can reduce the cost of search, accelerate synthesis, and support the recombination of heterogeneous information. Nevertheless, this finding should not be interpreted as evidence that GenAI itself possesses absorptive capacity. Rather, GenAI may expand the volume and variety of knowledge inputs available to firms, while AC remains the organizational capability through which these inputs are interpreted, validated, and applied. The relationship between AC and OC is positive, but weaker than the direct GenAI-OC association. This result is theoretically important because it qualifies the assumption that absorptive capacity automatically translates into creativity in AI-enabled environments. AC remains relevant because creative outcomes require more than access to ideas or information. Firms must select, evaluate, contextualize, and transform knowledge into ideas that are both original and useful. This interpretation is consistent with Mu-sa and Enggarsyah (2025), who show that knowledge transformation capabilities support innovation, agility, and resilience. It also aligns with Magis-tretti et al. (2021) and Cui (2025), who emphasize the role of AC in renewing the cognitive base of the firm and integrating external knowledge into innovation processes. However, the modest AC-OC effect observed in this study suggests that AC is better understood as an enabling condition rather than as a sufficient driver of OC. Its creative contribution is likely to depend on complementary factors such as leadership, domain expertise, psychological safety, organizational culture, and the quality of human-AI collaboration. The direct association between GenAI use and OC is the most salient result of the structural model. It suggests that generative technologies may contribute to organizational creativity by widening the range of available ideas, accelerating ideation cycles, reducing cognitive fixation, and facilitating the exploration of alternative solutions. This finding is consistent with Hörauf and Brem (2024), who show that AI can support idea originality by enabling novel semantic associations and broadening opportunities for conceptual recombination. At the firm level, the result indicates that GenAI may function as a creative catalyst by increasing the speed, diversity, and accessibility of ideational inputs. However, this direct association should not be interpreted in a technologically deterministic manner. GenAI may enrich the creative process, but the originality, relevance, and usefulness of the resulting ideas still depend on human judgment, organizational interpretation, and the ability to integrate AI-generated suggestions into meaningful work practices. The mediation result provides a more balanced understanding of the GenAI-AC-OC relationship. The indirect effect of GenAI use on OC through AC is positive but weak, indicating partial mediation. This means that AC explains part of the relationship between GenAI use and OC, but not its dominant share. Theoretically, this finding prevents an overly nar-row capability-based interpretation. GenAI does not appear to contribute to OC only because it strengthens AC. It also has a more immediate association with creativity through ideation support, experimentation, and exploratory search. The findings therefore support a dual-pathway view of GenAI-enabled creativity: a dominant direct augmentation pathway and a weaker, but still meaningful, capability-based pathway operating through AC. This dual interpretation helps explain why similar GenAI investments may produce different creative outcomes across firms. Access to generative tools can broaden ideational possibilities, but the conversion of these possibilities into organizational creativity depends on the routines through which firms assess relevance, detect limitations, connect AI-generated outputs with internal expertise, and embed validated ideas into organizational processes. In this sense, the findings support a hybrid creativity view in which GenAI enriches ideation while AC helps transform AI-mediated knowledge into more structured, contextualized, and actionable creative outcomes. The positive but weaker mediation effect suggests that this capability pathway exists, but that it must be reinforced through complementary managerial practices if firms wish to move beyond short-term ideation gains toward more sustained creative capability. Toward an integrative organizational model based on Human–AI collaboration The findings support the relevance of an integrative organizational model in which creativity emerges from the structured interaction between human competencies, knowledge-processing routines, and the generative capabilities of AI systems. The explanatory power observed for OC indicates that the proposed model captures an important part of the organizational conditions under which creativity develops in AI-enabled settings. However, the corrected structural results require a balanced interpretation. GenAI use is more strongly associated with OC through the direct pathway than through the indirect pathway mediated by AC. This suggests that GenAI contributes to creativity primarily by expanding ideational possibilities, accelerating exploratory search, supporting alternative framing, and facilitating iterative refinement. AC remains relevant, but its role is better understood as a complementary knowledge-processing capability through which firms evaluate, contextualize, recombine, and exploit AI-mediated outputs. This interpretation supports a hybrid creativity perspective. OC does not depend solely on individual ingenuity, nor does it emerge mechanical-ly from the adoption of generative technologies. Rather, it arises from the organization’s ability to orchestrate complementarities between human judgment, AI-generated suggestions, and routines for knowledge transformation. This view is consistent with Kolbjørnsrud (2024), who emphasizes the need to design hybrid organizations in which human and artificial intelligence are combined to enhance decision-making, learning, and innovation. In the present study, this logic is reflected in the dual role of GenAI: a dominant direct pathway associated with ideation and exploration and a positive but weaker indirect pathway operating through AC. This dual pathway is important because it avoids two reductive inter-pretations of GenAI-enabled creativity. A purely technological interpretation would overstate the autonomous creative power of GenAI and neglect the organizational conditions through which AI-generated outputs become meaningful. Conversely, a purely capability-based interpretation would underestimate the more immediate creative affordances of GenAI, particularly its capacity to broaden the search space, generate alternative framings, and accelerate experimentation. The findings therefore suggest that GenAI-enabled creativity is best understood as an organizationally situated process in which technological augmentation and knowledge absorption oper-ate jointly, although not with equal strength. Deliberate organization, rather than spontaneous emergence, is the broader implication of human-AI collaboration. When firms design routines that allow questioning, evaluating, contextualizing, and integrating AI-generated outputs into human decision-making and knowledge processes, GenAI can support OC. This interpretation is aligned with Herrmann and Pfeiffer (2022), who argue that the creative value of AI integration depends on the active shaping of sociotechnical interactions. The present study builds on this argument by demonstrating that AC serves as one organizational capability that can transform hybrid interaction into creative outcomes. Yet the comparatively weaker mediation effect suggests that AC alone is not sufficient. Its creative contribution is likely to depend on complementary organizational conditions, including leadership support, psychological safety, domain expertise, and governance mechanisms that preserve critical human oversight. Although the findings indicate positive associations between GenAI use, AC, and OC, they should not be interpreted as evidence that GenAI automatically enhances creativity. A growing body of research warns that intensive reliance on AI may generate cognitive dependence, reduce human creative effort, and weaken users’ creative skills over time. This caution is relevant because the positive direct association between GenAI and OC may coexist with longer-term risks that are not fully captured in a cross-sectional model. Crowston and Bolici (2025), for example, argue that excessive delegation of ideation tasks to generative models can reduce cognitive engagement and contribute to de-skilling. Similarly, Shukla et al. (2026) suggest that AI-based assistance may encourage cognitive offloading, while Cabrero-Daniel (2025) shows that prolonged exposure to AI-generated suggestions may produce anchoring effects that constrain intrinsic creativity. These insights qualify the interpretation of the findings by indicating that the creative value of GenAI depends on balanced human-AI collaboration. If organizations rely excessive-ly on generative systems without preserving human judgment, critical evaluation, and creative autonomy, the same technologies that support short-term ideation may weaken the endogenous capabilities required for sustained creativity. Theoretical and managerial implications The findings refine current understanding of the relationship between GenAI, organizational capabilities, and OC by showing that the creative value of generative technologies cannot be reduced either to technological adoption alone or to absorptive capacity alone. The corrected results indicate that GenAI use is strongly and directly associated with OC, while AC provides a positive but weaker mediating mechanism. Theoretically, this pattern supports a dual-pathway explanation of GenAI-enabled creativity. GenAI may directly enrich creative work by expanding the range of ideas, accelerating experimentation, and supporting exploratory problem-solving. At the same time, AC helps firms transform AI-mediated outputs into organizationally meaningful knowledge by enabling interpretation, evaluation, recombination, and exploitation. This result strengthens OC theory by extending the locus of creativity beyond individual cognition, organizational climate, and internal cultural conditions. These elements remain important, but in AI-enabled environments, OC increasingly depends on the interaction between human judgment, organizational routines, and intelligent technological artifacts. GenAI does not replace human creativity, nor does it independently generate organizational creativity. Its contribution lies in expanding the repertoire of possible ideas and enriching the informational base available to organizational actors. These possibilities become creatively valuable only when organizations possess the routines needed to assess their relevance, contextual fit, originality, and practical usefulness. The study also contributes to the AC perspective by showing that AC remains theoretically relevant in the context of generative technologies, even if its mediating role is not dominant. The strong association between GenAI use and AC suggests that generative technologies may enlarge the volume, variety, and accessibility of knowledge available to firms. However, the weaker AC-OC path indicates that knowledge-processing capability does not automatically translate into stronger creativity. To transform AI-generated insights into usable creative outcomes, organizations must combine informational abundance with interpretive discipline, domain expertise, and conducive organizational conditions. AC should therefore be understood as an enabling but not exhaustive mechanism in the GenAI-OC relationship. For managers, the findings imply that GenAI adoption should not be treated as a stand-alone technological investment. The direct association between GenAI use and OC suggests that firms may obtain creative benefits when employees use generative tools to explore alternatives, develop prototypes, compare ideas, and refine solutions. However, the organization is unlikely to sustain these benefits without routines for evaluating and integrating AI-mediated knowledge. Managers should therefore complement access to GenAI tools with structured practices for knowledge sharing, cross-functional interpretation, experimentation, and critical review. The results further indicate that managers should strengthen the conditions that support AC, even if AC plays a weaker mediating role than the direct GenAI-OC relationship. While GenAI enhances the speed and scale of information and suggestions, it does not dictate their quality, relevance, or originality. Organizations should therefore establish governance mechanisms that clarify how AI-generated outputs are assessed, how human expertise intervenes in validation, and how promising ideas are incorpo-rated into operational and strategic routines. Without such mechanisms, firms may increase the quantity of AI-generated content without improving the quality of creative outcomes. The findings also highlight the importance of organizational culture in supporting hybrid creativity. A culture that encourages experimentation, tolerates intelligent failure, values diverse perspectives, and supports cross-functional dialogue is more likely to transform GenAI outputs into creative value. In contrast, organizations that use GenAI passively may reproduce standardized outputs, reinforce existing cognitive frames, or become overly dependent on algorithmic suggestions. Therefore, the creative potential of GenAI hinges on whether organizations encourage employees to question, reinterpret, and recombine AI-generated ideas instead of merely adopting them. Finally, the study underlines the need for continuous capability development. Employees require not only technical familiarity with GenAI tools but also the analytical competencies needed to formulate effective prompts, evaluate the reliability of AI-generated content, identify bias or superficial novelty, and integrate machine-generated suggestions with human expertise. Therefore, GenAI is unlikely to replace human creativity in the future of OC. It will depend instead on the disciplined orchestration of human expertise, AC, and generative technologies within organizational routines capable of transforming AI-mediated knowledge into original and useful outcomes. Conclusions. This study examined the relationship between GenAI, AC, and OC in Moroccan firms. The findings indicate that GenAI use is positively associated with OC, both directly and indirectly through AC. This result suggests that GenAI should not be regarded merely as a technological instrument that accelerates ideation or supports experimentation. Rather, it should be understood as a knowledge-enabling resource whose creative value depends on the organization’s ability to acquire, interpret, recombine, and exploit AI-mediated knowledge. From this perspective, the study provides empirical support for a capability-based explanation of GenAI-enabled creativity: firms appear more likely to derive creative value from GenAI when they possess organizational routines that allow them to transform AI-generated outputs into meaningful knowledge and actionable ideas. The central contribution of the study lies in clarifying the mechanism through which GenAI use is associated with OC at the firm level. Whereas prior research has often emphasized the individual, task-level, or experimental effects of GenAI, particularly its capacity to support idea generation, reduce ideation time, and improve output quality, this study shifts the analytical focus toward organizational capability development. By positioning AC as a mediating capability, the findings show that the creative value of GenAI is neither automatic nor reducible to access to advanced digital tools. It depends on whether firms are able to evaluate AI-generated outputs critically, combine them with existing expertise, and embed them within organizational routines, decision processes, and innovation activities. This argument strengthens the theoretical connection between GenAI research, AC theory, and OC. Theoretically, the study advances the literature in several interrelated ways. It extends GenAI research beyond individual-level creativity and experimental settings by examining OC as a firm-level outcome. It also enriches AC scholarship by demonstrating the relevance of knowledge acquisition, assimilation, transformation, and exploitation processes in the context of generative technologies. In addition, it contributes to OC research by showing that creativity in AI-enabled environments is shaped by the interaction between technological resources and organizational learning capabilities. This perspective provides a more refined understanding of how firms may convert the informational and generative potential of GenAI into ideas that are both original and useful. From a management perspective, the results warn against seeing GenAI adoption as just a technical investment. Organizations seeking to enhance OC through GenAI should complement technological deployment with learning routines, knowledge-sharing practices, employee training, and governance mechanisms that support the critical evaluation of AI-generated outputs. Managers should pay particular attention to employees’ ability to formulate effective prompts, assess the relevance and reliability of AI suggestions, and integrate these suggestions with contextual knowledge and strategic priorities. In this respect, AC becomes a central managerial lever for transforming GenAI from an experimental tool into a source of creative and organizational value. Several limitations should be acknowledged. The cross-sectional design restricts the possibility of making strong causal claims; therefore, the relationships identified in the model should be interpreted as theoretically grounded associations rather than definitive causal effects. The reliance on self-reported data from single respondents may also raise concerns regard-ing common method bias, although procedural and statistical precautions were considered. Furthermore, the Moroccan empirical context provides valuable evidence from an emerging economy, but it may limit the direct generalizability of the findings to other institutional and technological environments. Future research could address these limitations by employing longitudinal designs, multi-informant data, archival or behavioral indicators of creativity, and cross-country comparisons. Further studies could also examine whether governance practices, digital maturity, organizational culture, or human-AI collaboration skills moderate the relationship between GenAI, AC, and OC. In conclusion, this study indicates that the creative value of GenAI depends not only on the availability of generative technologies but also on the organizational capabilities through which firms absorb, evaluate, and apply AI-mediated knowledge. This insight is particularly relevant in emerg-ing-market contexts, where firms may adopt advanced digital tools without necessarily possessing the routines required to transform them into sustained creative outcomes. By identifying AC as a central explanatory mechanism, the study offers a more nuanced account of how GenAI can support OC and provides a foundation for future research on AI-enabled capability development. Acknowledgements. The authors would like to thank the anonymous reviewers and the editorial team for their constructive comments and valuable suggestions, which helped improve the quality, clarity, and methodological rigor of the manuscript. The authors also acknowledge the participating firms and respondents, whose time, cooperation, and professional insights made the empirical part of this study possible. AI declaration statement Intelligence Policy of the publisher. Compliance with ethical standards This article does not contain any studies with human participants or animals performed by the authors. Extracting and inspecting publicly accessible files (scholarly sources) as evidence, before the research began no institutional ethics approval was required. Data availability statement All data generated or analyzed are included in the published article. The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation. The raw anonymized data can be provided by emailing the primary author. All listed authors have made a substantial, direct and intellectual contribution to the work, and approved it for publication. The authors take full responsibility for the accuracy and the integrity of the source analysis. The authors declare that they got acquired, accept and apply General Artificial The authors declare that they have not used AI tools for research process. Author contributions Conflict of interest statement cial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. The authors declare that the research was conducted in the absence of any commer- Annex Appendix 1 Survey Description: Appendix 1 presents the structured survey instrument used for data collection. The survey was designed to capture both firm-level profile information and respondents’ assessments of the three main constructs of the study: GenAI use, absorptive capacity, and organizational creativity. The first part collects descriptive information on the firm, including a sector of activity, annual revenue, number of employees, and firm age. The second part includes the measurement items used to assess the extent to which GenAI is integrated into organizational activities, the firm’s capacity to acquire, assimilate, transform, and exploit knowledge, and its ability to generate and support novel and useful ideas. All construct-related items were measured using a five-point Likert scale ranging from 1, “strongly disagree,” to 5, “strongly agree.” This structure ensures consistency across responses and allows the survey data to be used for the subsequent PLS-SEM analysis. Instruction: Please tick the appropriate answer for the company profile and indicate your level of agreement with each statement in the main questionnaire. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 3 Employees regularly use generative AI tools to improve their productivity. Use of Generative AI ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 4 Our organization considers generative AI a strategic tool for value creation. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 5 The use of generative AI significantly improves the quality of internal outputs. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 1. Our firm quickly identifies relevant new external knowledge. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 2 Our firm has the capability to analyze and understand new technologies. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 3 Our firm readily integrates new knowledge into its internal processes. Absorptive Capacity ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 4 Our firm is able to transform new knowledge into concrete opportunities. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 5 Our firm effectively exploits new knowledge to improve performance. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 1 Our organization actively encourages the generation of new ideas. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 2 Employees often develop original solutions to problems. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 Organizational Creativity 3 Our organization regularly develops new concepts, products, services, or practices. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5 4 The work environment fosters experimentation and initiative-taking. ☐ 1 ☐ 2 ☐ 3 ☐ 4 ☐ 5