1. Introduction
AI leads to worldwide industry transformations by providing major enhancements for decision processes together with operational performance and innovative capabilities. The rapid growth of AI implementation across sectors presents difficulties for its workflow integration into performance maintenance, especially in developing economic systems. The backbone of Pakistan’s economy consists of SMEs that encounter special obstacles to AI implementation because they lack money and have poor infrastructure and lack experienced workers [
1,
2]. Manufacturing firms in Pakistan, especially SMEs, must focus on technological readiness because it determines both AI integration success and performance sustainability [
3,
4] in their less-developed market. Pakistan’s manufacturing SME sector stands as its top export contributor because it serves as the perfect environment to understand how AI adoption along with technological readiness creates sustainability when resources remain limited [
3,
4].
Some aspects might be technological readiness and adjustment of AI for SMEs in Pakistan, which are the most important for survival and competitive advantage in rapidly changing emerging market [
5]. The advancement of AI throughout various industries globally has not triggered Pakistani SMEs to progress since they face structural barriers including insufficient financial resources, limited skilled labor, and poor infrastructure. The stagnation does not affect their essential role in the economic landscape. Pakistan’s economy relies heavily on SMEs, which make up 90% of all enterprises while producing 40% of GDP and accounting for 25% of total exports. These statistics establish their vital position within Pakistan’s economic structure, especially in export sectors that include textiles along with surgical instruments and sports goods. The slow adoption of global technology requires immediate investigation because Pakistani SMEs cannot match worldwide technological progress [
3,
4]. AI presents substantial advantages for Pakistani SMEs, but their full utilization remains restricted by employee technological skills shortages, inadequate training programs, and limited funding sources. The continuous commitment to training programs becomes indispensable for Pakistani SMEs that need to bridge their skill gaps and make their personnel ready to use artificial intelligence effectively. The research findings demonstrate that continuous capacity development helps organizations become more adaptable, which enables SMEs to use AI for long-term performance success in changing markets [
6,
7]. To effectively navigate the challenges associated with AI adoption, SMEs must go beyond mere technological acquisition and focus on the continuous development of their internal capabilities. According to the dynamic capabilities theory (DCT) [
8], organizations must consistently adapt, integrate, and reconfigure their resources to maintain a competitive edge in a rapidly evolving landscape, and organizational learning serves as a crucial capability that enables SMEs to overcome knowledge gaps and effectively integrate AI into their operations despite resource constraints. However, learning alone is insufficient; SMEs must complement it with financial investment, collaboration, and supportive policies to fully embed AI into their business processes [
2]. The implementation of AI systems depends on well-structured data management, a challenge many resource-limited Pakistani SMEs struggle to address. By leveraging DCT principles, SMEs can enhance their technological readiness through structured learning mechanisms that facilitate knowledge acquisition and resource reconfiguration. These learning systems help develop essential data-processing capabilities, allowing SMEs to optimize AI adoption and fully capitalize on its potential [
2,
6]. Furthermore, capacity-building initiatives that focus on both technical and analytical skills ensure that employees can effectively interpret AI-generated insights and apply them strategically to improve performance. This aligns with [
8]’s argument that firms must dynamically adjust their competencies to sustain a competitive advantage in volatile markets.
While AI adoption and technological readiness have been recognized as key drivers of organizational success, existing research tends to focus on these factors in isolation, often emphasizing their direct impact on organizational performance [
9,
10,
11]. Studies on AI adoption have highlighted its potential to enhance operational efficiencies and decision-making capabilities, but they frequently overlook the critical role of technological readiness in maximizing the benefits of AI [
12,
13]. Technological readiness, which reflects an organization’s capacity to integrate and leverage new technologies, is essential for AI’s successful implementation but remains underexplored in its joint effect with AI adoption. Firms can rapidly adapt to changing technological conditions through organizational learning, which establishes itself as their primary dynamic capability. Technological readiness enables employees to receive regular skill and knowledge updates so they can properly work with AI and advanced technological systems [
14,
15]. The organizational learning process creates an experimental culture for sharing knowledge that becomes vital for operational process integration of technological readiness [
16]. The moderating function plays a vital part in developing-economy small and medium-sized enterprises because they need to extract maximum value from their technology investments. According to the DCT, a company achieves a lasting competitive advantage by understanding that it must adapt its resources alongside integrating them and reconfiguring them to adjust to shifting market environments [
8]. Organizational learning functions as a capability mobilizer which helps businesses transform their technology transitions and develop employee capabilities [
17]. Past research [
6] has shown direct connections between AI adoption and technological readiness and performance results, but insufficient examination exists regarding organizational learning’s ability to boost employee capacity development during technological integration. The lack of research becomes noticeable in Pakistani SMEs because these firms must use organizational learning strategies to extract maximum value from AI technologies [
14,
15].
The existing literature tends to treat these variables independently, leaving a significant gap in understanding how technological integration, coupled with organizational learning, fosters employee capacity development and translates into improved organizational performance [
7]. Additionally, the mediating role of employee capacity building, as a dynamic capability that facilitates the effective use of AI and other technologies, has not been sufficiently addressed. Current research largely overlooks the mechanisms by which technological integration enhances human capital and, in turn, organizational performance. This study fills this gap by examining the integrated effects of technological readiness and AI adoption, moderated by organizational learning and mediated by employee capacity building, thus offering a more comprehensive framework for understanding how organizations can leverage technological integration to achieve sustained performance improvements [
6]. Therefore, this study has the following three objectives:
RO1: To assess the role of implementation of technological readiness and AI adoption in sustainable performance of SMEs.
RO2: To compute the mediating role of employee capacity building between technological readiness, AI adoption, and sustainable performance of SMEs.
RO3: To analyze the moderating role of organizational learning between technological readiness, AI adoption, and employee capacity building.
This study makes three key contributions to the existing literature. First, this study pioneers the integration of AI adoption and technological readiness within the DCT framework, extending its application to resource-constrained SMEs in developing economies like Pakistan. Second, it identifies employee capacity building as the critical mediator through which technological readiness and AI adoption translate into sustainable performance, addressing a gap in prior research that overlooked human capital’s role in technology-driven sustainability. Third, it reveals organizational learning as a boundary condition that amplifies the impact of technological readiness but not that of AI adoption, offering nuanced insights into how adaptive learning cultures uniquely enable SMEs to optimize resource-limited technological transitions.
In the following section, we outline the theoretical underpinnings and hypothesized relationships. We then detail the methodological approach adopted for our research and share findings from the empirical analysis. Subsequently, we delve into the implications of our study, both from a theoretical standpoint and in terms of practical application. Finally, we reflect on the study’s limitations and suggest directions for future research in this area of study.
3. Materials and Methods
3.1. Samples
As the primary aim of the study is to establish the influential relationship between variables, employing a quantitative research approach is most advantageous [
67]. In order to assess our hypothesis, we conducted a survey to collect data on latent constructs utilized in the research framework. The target population for this study was exporting firms authorized by the Securities and Exchange Commission of Pakistan (SECP). The exporting firms were chosen due to their substantial contribution to Pakistan’s economy. The research concentrated on exporting firms because they play a vital economic function by driving both GDP growth and employment generation in Pakistan. Global markets pressure export-oriented SMEs to integrate AI into their operations since innovation combined with efficiency represents essential elements for staying competitive overseas. The research environment provides strong grounds for understanding the relationships between AI adoption, technological readiness, and employee capacity development for sustainable performance outcomes. The research acknowledges possible self-selection bias as a potential weakness from studying exporting firms exclusively. We selected this research approach to track technological competitive dynamics in international markets, but this decision restricts the universal applicability of research results to firms that limit their business activities to the domestic market. Research should expand to include non-exporting SMEs to validate if the established relationships exist similarly in less globally exposed business environments. In this scenario, it is very important to overview the exporting sector of developing countries like Pakistan to ensure they are ready to accept these rapid changes. Consequently, Pakistani exporting firms were selected as the research sample for this study. Data collection occurred from August to October 2024, and the participants were selected using convenience sampling, targeting managerial-level employees in export-oriented SMEs across five major Pakistani cities (Karachi, Lahore, Sialkot, Faisalabad, and Gujranwala). This sampling technique was chosen for its practicality and ability to provide access to a suitable sample within the constraints of time and resources. The managerial-level employees were directly involved in decision-making processes and AI adoption, making them ideal respondents. While probabilistic methods could enhance generalizability, they were not feasible given logistical constraints and the specific focus on SMEs. Combining methods was avoided due to the complexity and resource limitations inherent in the study’s design. The potential limitations of convenience sampling, including reduced generalizability, have been acknowledged and discussed [
68].
We distributed 750 questionnaires to higher-level managers, with official authorization, inviting their participation in data collection because of their familiarity with organizational policies and procedures. As a result, 495 questionnaires were returned; after careful scrutiny, 15 responses did not meet the standards, i.e., having double-ticked on a response. After excluding these incomplete responses, we obtained 480 valid questionnaires, yielding a response rate of 66 percent. The survey instrument underwent a rigorous validation process to ensure its reliability and relevance. Initially, we conducted consultations with three academic experts in AI and SME management to refine the questionnaire items. This was followed by a pilot study involving 30 participants from diverse SMEs to assess the clarity and applicability of the questions. Feedback from the pilot study was used to make necessary modifications, enhancing the instrument’s validity and reliability. The theoretical underpinnings guiding the operationalization of concepts were based on DCT and validated scales from prior studies [
16,
69].
The survey achieved a sample size of 480 managerial-level employees from export-oriented SMEs in Pakistan. This sample size was selected to ensure sufficient statistical power and representativeness. To evaluate the reliability of the survey results, a 95% confidence level was used. Assuming a population size of approximately 500,000 SMEs in Pakistan, with an estimated margin of error of ±4.5%, the sample size meets the recommended thresholds for survey research. The 95% confidence interval implies that if the study were repeated 100 times with different samples from the same population, 95 of those samples would produce results within the calculated range. This statistical consideration strengthens the reliability of our findings, although some caution is warranted due to the convenience sampling method. Despite this limitation, the diversity of industries and organizational roles in the sample contributes to its robustness and applicability to the broader SME population.
Table 2 provides the detailed demographic information of the respondents and exporting firms.
3.2. Measures
The ultimate questionnaire utilized in this study was crafted through three sequential steps, as [
70] recommended. Initially, scales for all variables were established through a comprehensive review of the pertinent literature. Subsequently, after consultations with three scholars related to the field, the measuring scale underwent revisions to enhance clarity and comprehensibility. The structured questionnaire consisted of five main sections, each targeting a specific construct: AI adoption, technological readiness, employee capacity building, organizational learning, and sustainable performance. Items were measured using a five-point Likert scale (1 = strongly disagree to 5 = strongly agree), ensuring consistency and ease of response for participants. The selection of these measurement scales was guided by established theoretical frameworks, including DCT and validated scales from prior studies [
16,
69]. Specific details of measures are as follows:
Artificial Intelligence (AI):
AI adoption was measured using five items adapted from the validated scale by [
71]. The items captured various dimensions of AI adoption, including the use of AI technologies for decision-making, automation, and innovation. Participants responded on a five-point Likert scale (1 = strongly disagree to 5 = strongly agree). Sample items include the following:
‘Our organization uses AI technologies for operational optimization’.
‘AI is integrated into our decision-making processes’.
The scale demonstrated strong reliability in prior studies (Cronbach’s α = 0.80) and was confirmed to be reliable in this study (Cronbach’s α = 0.77).
Technological Readiness:
Technological readiness was measured using five items adapted from the Technology Readiness Index (TRI) developed by [
69]. The scale assessed the preparedness of organizations to adopt and integrate new technologies. Sample items include the following:
‘Our organization has the technical infrastructure to implement advanced technologies’.
‘Our employees are open to learning about new technological systems’.
Responses were recorded on a five-point Likert scale, and the scale exhibited strong reliability in this study (Cronbach’s α = 0.78).
Employee Capacity Building:
Employee capacity building was measured as a mediating variable using five items adapted from the scale developed by [
16]. The scale assessed training, skill enhancement, and knowledge-sharing practices within the organization. Sample items include the following:
‘Our organization provides regular training programs to enhance employee skills’.
‘Employees are encouraged to acquire knowledge about new technologies’.
Responses were recorded on a five-point Likert scale, and the scale showed high reliability (Cronbach’s α = 0.86).
Organizational Learning:
Organizational learning, as the moderating variable, was measured using four items adapted from [
72]. The scale assessed the extent to which the organization fosters knowledge-sharing, experimentation, and continuous learning. Sample items include the following:
‘Our organization encourages employees to learn from past experiences’.
‘Knowledge-sharing across departments is a common practice in our organization’.
Responses were recorded on a five-point Likert scale, and the scale demonstrated high reliability (Cronbach’s α = 0.88).
Sustainable Performance:
Organizational performance was measured using four items adapted from [
73]. The scale captured operational efficiency, innovation, and market competitiveness. Sample items include the following:
‘Our organization has improved its operational performance over the past year’.
‘We are consistently meeting our performance targets’.
Responses were recorded on a five-point Likert scale, and the scale exhibited strong reliability (Cronbach’s α = 0.79).
3.3. Common Method Bias
To address potential common method bias (CMB), we employed Herman’s single-factor test and the marker variable technique [
74]. Herman’s test indicated that the first factor explained only 31% of the variance, which is well below the threshold of 50%, suggesting minimal risk of CMB [
75]. Additionally, the marker variable technique was implemented by introducing a theoretically unrelated variable to the model. The correlations between this marker variable and the primary constructs were found to be negligible, reinforcing the conclusion that CMB does not significantly influence the study’s results [
76,
77]. These combined methods provide robust evidence for the reliability and validity of our findings.
4. Results
In this study, we employed partial least squares structural equation modeling (PLS-SEM) for data analysis, as it is particularly suitable for studies involving complex models, smaller sample sizes, and exploratory research contexts. PLS-SEM evaluates both measurement models and structural models, enabling simultaneous assessment of construct validity and hypothesized relationships. Compared to covariance-based SEM, PLS-SEM is more robust in handling non-normal data distributions and smaller samples, making it an ideal choice for this study. This approach provided comprehensive insights into the relationships between AI adoption, technological readiness, employee capacity building, and sustainable performance [
78,
79]. It is possible to model both formative and reflective constructs in PLS-SEM (4.1); the software is suitable for research with small sample sizes and has no problems with multicollinearity. Also, user-friendly software and clear guidelines for the assessment of measurement and structural models enhance the application of this tool in the investigation of consumer behavior, organizational behavior, education, health psychology, and social psychology [
80]. In this context, the application of the PLS-SEM approach through Smart PLS 4 is exceptionally reasonable because of the availability of numerous options for model assessment, the software’s intuitive interface, and the powerful data visualization tools that help to provide a detailed analysis and a clear interpretation of the results, thus providing a high level of research relevance and significance.
4.1. Measurement Model
The reliability and validity results in
Table 3 and
Figure 2 confirm robust psychometric properties for all constructs, with item loadings > 0.60, Cronbach’s α > 0.70, composite reliability > 0.70, and AVE > 0.50, ensuring internal consistency and discriminant validity [
80,
81]. However, the descriptive statistics reveal high mean scores for key variables (e.g., AI adoption: 3.88–4.26; employee capacity building: 3.69–4.23), suggesting sampled export-oriented SMEs exhibit above-average technological maturity. The limited variability (SD: 0.81–1.19) likely stems from the study’s focus on export-focused firms, which often prioritize technological readiness and employee training to meet international standards. While this homogeneity underscores the sample’s alignment with global competitiveness, it may constrain generalizability to domestic SMEs with lower resource access.
The high means align with Pakistan’s export sector dynamics, where SMEs face pressure to adopt advanced technologies for market survival. Nevertheless, PLS-SEM remains critical for unpacking the structural relationships between constructs (e.g., mediation/moderation effects), which descriptive statistics alone cannot capture. Future research should incorporate non-export SMEs to enhance variability and validate findings across diverse contexts. Despite sample-specific trends, the model’s predictive relevance (Q2 >0) confirms its utility in explaining how AI and readiness strategies interact to drive sustainable performance, even in high-performing cohorts.
The discriminant validity of the constructs was checked by the Hatrotrate–monotrate (HTMT) ratio and the Fornell–Larcker criterion, which is presented in
Table 4. The HTMT of all the construct pairs was less than 0. 85, which clearly depicts that these two variables are different and measure different attributes of the theoretical model. Discriminant validity was further tested using the Fornell–Larcker criterion, which revealed that the square root of AVE for each of the constructs was higher than the maximum coefficient correlation between the construct and all the other constructs (
Table 5). These results support that the constructs are different theoretical concepts, thus enhancing the measurement model’s validity [
82].
4.2. Predictive Relevance
Table 6 shows the predictive relevance of the two independent variables, ECB and SP, in the research model. The R-square and the adjusted R-square are significant measures that help understand the proportion of the total variation accounted for by the predictors in the model. For ECB, the value of the R-square is 0. 583, and for the adjusted R-squared, it is 0.592, meaning that the proportion of the change in the predictor is due to ECB. Correspondingly, for SP, the R-square value is equal to 0.557, and for the adjusted R-square value, it is equal to 0.563; this means that the model explains 55.7 percent of the variation in the dependent variable. The results indicate that SP is well explained by the model, with an R
2 value of 0.557, meaning that the independent variables account for 55.7% of its variance. The strong predictive power of ECB further reinforces its critical role in influencing SP. Moreover, the extent of predictive relevance of the model was tested further by applying the blindfolding procedure, particularly, the Stone–Geisser Q
2. The predictive relevance of latent outcome constructs has been estimated based on the Stone–Geisser Q
2 using PLS blindfolding. The Q
2 statistic represents a measure of how well the observed values are reconstructed by the model and its parameter estimates. A Q
2 value greater than zero indicates that the model has predictive relevance [
83]. In our analysis, the Q
2 value for ECB is 0.352, derived from the Sum of Squares Total (SSO) of 2400.000 and Sum of Squares Error (SSE) of 1854.279, suggesting substantial predictive relevance. Similarly, the Q
2 value for SP is 0.243, with an SSO of 1920.000 and an SSE of 1657.204, indicating moderate predictive relevance.
Following the recommendations of [
78,
84],
Table 7 and
Figure 3 depicts the varying magnitudes of the coefficients in H1 (β = 0.228, AIA → ECB) and H2 (β = 0.371, TR → ECB), reflecting distinct roles of technological integration in employee capacity building. The coefficients reveal that TR (β = 0.371) has nearly double the economic impact on ECB compared to AIA (β = 0.228). This aligns with DCT’s emphasis on TR as a foundational capability, where every unit increase in TR yields a 16.3% greater workforce competency improvement than AIA. For SMEs, prioritizing TR investments (e.g., infrastructure and training) offers higher marginal returns. TR demonstrates a stronger direct impact, as it establishes foundational infrastructure and cultural adaptability, enabling SMEs to systematically upskill employees and align workflows with emerging technologies. In contrast, AIA requires specialized expertise and faces implementation barriers (e.g., data literacy gaps), resulting in a comparatively smaller effect. This aligns with DCT; TR fosters adaptive capacity (resource reconfiguration), while AI adoption demands complementary learning mechanisms to translate tools into skills.
Similarly, the robust effect of ECB on sustainable performance (H3: β = 0.497) underscores that human capital development is pivotal for leveraging technological investments into tangible outcomes. Following the recommendations of [
85], the mediation effects (H4/H5) further emphasize that AI and TR indirectly enhance performance through ECB, rather than acting in isolation.
The above results show that export-oriented Pakistani SMEs likely exhibit higher baseline TR (mean = 4.0), amplifying its direct impact. AI adoption, while significant, remains constrained by skill mismatches, explaining its smaller coefficient. Future research should explore how structured learning interventions could amplify AI’s role in ECB.
4.3. Mediation Analysis
The mediation analysis reveals distinct pathways through which TR and AIA influence SP via employee ECB. While both TR → ECB → SP (β = 0.184, p < 0.001) and AIA → ECB → SP (β = 0.113, p < 0.001) are significant, the larger coefficient for TR underscores its foundational role in enabling SMEs to translate technological investments into sustainable outcomes. ECB’s mediation effect is stronger for TR → SP (β = 0.184) than AIA → SP (β = 0.113), indicating that TR’s systemic readiness amplifies human capital’s role in sustainability. A 10% TR improvement indirectly boosts SP by 1.84%, whereas the same AIA increase yields 1.13%, highlighting TR’s strategic precedence. This divergence arises because TR encompasses not only infrastructure but also cultural and strategic alignment, which collectively empower employees to assimilate advanced technologies effectively. In contrast, AI adoption, while impactful, depends heavily on pre-existing technological preparedness to realize its full potential. The smaller indirect effect of AIA aligns with Pakistan’s export-oriented SME context, where firms prioritize TR to meet global standards, yet they face skill gaps in deploying AI autonomously. This mirrors findings in developing economies, where TR often precedes AI’s value extraction. Despite the sample’s limited variability (high means: TR = 4.0, AIA = 3.9), the results emphasize that TR’s systemic readiness amplifies ECB’s mediating role, whereas AI adoption requires complementary training to bridge implementation barriers.
4.4. Moderation Analysis
The moderation results reveal critical insights into how OL differentially influences technological integration pathways. While OL significantly amplifies the relationship between TR and ECB (β = 0.187, p = 0.001), it does not enhance the direct link between AIA and sustainable SP (β = 0.042, p = 0.170). TR involves systemic infrastructure, cultural adaptability, and strategic alignment. OL strengthens this relationship by fostering knowledge-sharing and iterative learning, enabling SMEs to translate readiness into actionable employee skills (e.g., training programs tailored to new technologies). This aligns with DCT, where OL acts as a reconfiguring mechanism, helping firms adapt TR investments into human capital. Moreover, AI adoption often relies on predefined tools (e.g., automated workflows) that may not require deep organizational learning to yield immediate efficiency gains. In export-oriented SMEs, AI tools are frequently deployed for specific tasks (e.g., inventory management), where performance improvements depend more on technical implementation than adaptive learning. This mirrors findings in developing economies where AI’s standalone utility often precedes learning-driven refinements.
4.5. Discussion
The research results deliver essential knowledge about exporting SMEs operating in Pakistan. The nature of exporting firms exposes them to strong global market competition alongside fast-changing demands that compel them to adopt advanced technologies including AI to stay competitive. AI adoption together with technological readiness strengthens employee capacity development, which allows SMEs to fulfill international standards and enhance operational performance. The results demonstrate that capacity-building initiatives represent a vital sustainable performance strategy because they align with international market standards. The exclusive examination of exporting SMEs creates the risk of self-selection bias that should be considered. The existing technological readiness and adaptability of export-oriented firms might enhance the strength of observed relationships because these firms have already demonstrated superior readiness attributes. Future research needs to investigate whether the findings from exporting SMEs in Pakistan can apply to domestic-only SMEs because these businesses operate under various distinct market conditions.
In terms of AI adoption, this study found that while AI adoption has a positive and significant impact on employee capacity building in Pakistani SMEs, the effect size is relatively small compared to technological readiness. This suggests that AI alone does not automatically drive workforce development; rather, its effectiveness is contingent upon an organization’s ability to continuously adapt, integrate, and reconfigure its resources, as emphasized by DCT. According to DCT, firms must develop learning capabilities and resource reconfiguration mechanisms to fully leverage technological advancements. The study’s assertion that “resource reconfiguration underpins sustainable outcomes” directly supports our findings. Mexican manufacturing firms leveraging green supply chain practices mirror Pakistani SMEs’ reliance on TR and OL: both contexts show that systemic readiness (e.g., eco-friendly infrastructure) combined with learning-driven adaptability (e.g., employee training) are critical for translating technological investments into performance gains. This cross-context validation reinforces DCT’s universality in resource-constrained settings [
86]. In the case of SMEs, technological readiness serves as a foundational capability, enabling firms to effectively integrate AI into their operations and workforce development strategies [
30]. Additionally, the insignificant effect of AI in H6 suggests that AI adoption alone does not significantly enhance the relationship between organizational learning and employee capacity building. This finding reinforces the DCT perspective that technological integration must be accompanied by adaptive learning processes to create sustainable competitive advantages. Without strong learning mechanisms and structured training programs, SMEs may struggle to translate AI adoption into meaningful performance improvements. Therefore, while AI adoption can enhance operational efficiency, its role in fostering employee growth and sustainable organizational performance depends on an organization’s ability to reconfigure resources dynamically in response to technological change. Our findings align with DCT’s emphasis on adaptability and resource transformation, and SMEs that invest in capacity-building initiatives and organizational learning are better positioned to extract long-term benefits from AI adoption. This supports previous research [
15], which highlights the role of organizational learning in maximizing AI’s impact. Similarly, [
14] stresses the importance of innovation and experimentation, reinforcing our observation that technological readiness and employee capacity building serve as critical enablers of sustainable performance. Unlike prior studies, our research explicitly integrates AI adoption and technological readiness within the DCT framework. DCT emphasizes that for SMEs to achieve a sustained competitive advantage, they must continuously reconfigure their resources and capabilities to adapt to shifting technological landscapes. In the context of AI adoption, our findings highlight that simply adopting AI technology is not enough; firms must actively engage in dynamic capability processes, such as learning and resource reconfiguration, to translate AI adoption into long-term performance improvements. This is especially critical for SMEs operating in resource-constrained and rapidly changing environments, where the ability to continuously learn, adapt, and reconfigure capabilities is essential for leveraging AI effectively. By applying DCT, our study shows how organizational learning and capacity building serve as key mechanisms through which SMEs can adapt to AI adoption, bridging the gap between technological readiness and sustainable performance outcomes.
For technological readiness, this study found that technological readiness significantly enhances employee capacity building. This finding elaborates the importance of organizational preparedness in enhancing workforce skills. Prior studies have shown that technological readiness equips firms with the infrastructure and mindset needed to adopt new technologies, thereby enabling employees to build relevant competencies [
87]. In developing countries like Pakistan, technological readiness is more crucial because it provides employees with access to training and development tools, fostering skill development in response to technological shifts [
31]. The finding indicates that technological readiness acts as a catalyst for capacity building, ensuring that employees can adapt to and leverage technological advancements, ultimately contributing to improved organizational performance and sustainability [
23].
In terms of employee capacity building, this study found a positive impact of employee capacity building on sustainable performance of SMEs. This result indicates the importance of a skilled workforce in achieving long-term sustainability. Similar results have also been presented in prior studies that capacity building equips employees to implement sustainable practices, improve resource efficiency, and foster innovation [
34]. In resource-constrained environments like Pakistan, well-trained employees are pivotal in aligning operational processes with sustainability goals, leading to measurable improvements in performance [
11]. This finding underscores the importance of human capital in driving sustainable outcomes, as employees’ ability to effectively use technologies and innovate within their roles ensures comprehensive organizational performance [
9].
For mediating relationships, this study found a significant mediating role of employee capacity building between AI adoption as well as technological readiness and sustainable performance of SMEs. Previous studies suggest that while AI and technological readiness enhance operational capabilities, it is the development of employee skills that allows organizations to fully leverage these technologies for sustainable performance [
88]. Employee capacity building ensures that employees can apply AI and new technologies to improve resource efficiency, reduce waste, and integrate sustainability into business practices [
89]. These results demonstrate that human capital acts as a critical enabler, turning technological potential into comprehensive organizational success across economic, environmental, and social dimensions [
18].
With regard to moderating relationships, this study found an insignificant moderating relationship of organizational learning between AI adoption and employee capacity building but found a significant moderating relationship of organizational learning between technological readiness and sustainable performance. According to the results, the significant moderating role of organizational learning between technological readiness and employee capacity building underscores the importance of learning practices in helping employees adapt to new technologies, as supported by previous research [
21]. Organizational learning enhances employees’ ability to absorb and utilize technology, driving capacity building. However, the insignificant moderating effect of organizational learning between AI adoption and employee capacity building may be due to the fact that AI technologies are often user-friendly, meaning they typically require less extensive learning or training for employees to use them effectively. As a result, organizational learning may not play as significant a role in enhancing the relationship between AI adoption and employee capacity building in this case [
27]. Additionally, organizations may lack the structured learning systems needed to fully support AI adoption, limiting its interaction with capacity building [
35]. This suggests that while learning is crucial for readiness-driven technologies, AI may operate independently of deep learning cultures in this context.
4.6. Theoretical Implications
This study makes significant theoretical contributions by extending the application of the DCT to the context of SMEs in developing economies. While previous research has largely focused on large organizations in developed markets, our findings illustrate how resource-constrained SMEs can leverage dynamic capabilities, such as organizational learning, to enhance employee capacity and achieve sustainable performance [
90]. First, it confirms that AI adoption and technological readiness are positively related to employee capacity building, a relationship little discussed in the literature. This research fills a gap in how digital transformation enhances human capital development by directly linking technology adoption with human capital development. Second, this study contributes to the literature on sustainable performance, demonstrating employee capacity building as a significant mediator of sustainable performance in the relation between technological building blocks (AI adoption and technological readiness) and sustainable performance. While some studies have investigated direct relationships between technology and performance outcomes, little attention has been paid to the mediating role of human resource development [
1,
23].
Grounded in DCT, this research extends the DCT approach by uncovering employee capacity building as a dynamic capability that serves to mediate the relationship between technological integration and sustainable performance. The novel contribution of this study is its empirical examination of the mediating role of employee capacity building, showing that firms that build their workforces are in a stronger position to transfer technology investments into superior long-term sustainable performance. Additionally, this study extends DCT by testing the moderating effect of organizational learning and highlighting that organizational learning significantly amplifies the relationship between technological readiness and employee capacity building in the context of AI adoption. The argument that these technologies moderate learning strengths differentially deepens the theory that AI technologies might be less burdensome to integrate without major learning, whereas most other investments need a relatively more robust learning capability development to build capacity.
This study registers its contribution to the literature as one of the very first studies to visually depict the specific mediating role of employee capacity building in linking AI adoption and sustainable performance, and technological readiness and sustainable performance within the highly misunderstood context of SMEs. These results contribute strong evidence that employee capacity building is a crucial mechanism through which technological adoption translates into better organizational outcomes. Further, this study delineates the conditions under which learning actually adds to the benefits derived from investments in technology by discovering a significant moderating role for organizational learning in the relationship between technological readiness and employee capacity building. This study also reveals how insignificant the moderating effect of organizational learning on AI adoption and employee capacity building is, as the unique nature of AI technology suggests that formal learning is less important to effectively embed in organizational processes.
By demonstrating how technological integration is converted to sustained performance through employee empowerment, this study contributes to applying DCT in AI adoption and technological readiness. It notes that adoption of AI and technological readiness do not directly translate into performance but rather require organizations to build up and leverage the capabilities of their employees. Furthermore, the study adds depth to the discourse on organizational learning by revealing that technology aspects require a more structured approach to be beneficial, while AI adoption may, in fact, benefit from its user-friendly, autonomous nature.
4.7. Practical Implications
The research delivers practical recommendations to both SME managers and policy makers in developing countries such as Pakistan about improving organizational success through AI implementation and workforce development programs. Exporting SMEs should allocate funds to deliver specialized training that teaches their employees data analysis techniques combined with AI decision processes and machine learning fundamentals. SME managers need to develop a learning-focused organizational culture through programs that support knowledge exchange between departments together with ongoing training programs which help workers adjust to AI requirements [
24]. SMEs should integrate AI through affordable applications first to optimize their inventory and manage customer relationships while building capability for more advanced systems. Currency policy through tax relief along with subsidized training and technology grants from policymakers will strengthen the export capability of international SMEs. The adoption of these strategies helps SMEs break AI implementation barriers and develop technological readiness, which enables their sustainable market growth in changing business environments [
87].
Moreover, this research highlights the significance of aligning technological approaches with organizational learning and workforce growth for achieving sustainable performance. In addition, the strong correlation between technological readiness and employee capacity building indicates that SMEs should evaluate their existing technological capabilities and allocate resources to modernizing their infrastructure. Nonetheless, investing in technological tools is not enough as firm leaders need to cultivate an organizational environment that encourages continuous learning and development so that employees are well equipped to adopt and harness these new technologies. Furthermore, organizational learning increases the effects of technological readiness on AI adoption but showed little impact on AI adoption [
23]. This implies that AI technologies might represent lower effort organizational learnings for an effective implementation cycle, which might make them more accessible for SMEs with limited learning resources. Thus, management at a senior level must view AI as a strategic investment.
Managers of SMEs can implement multiple specific actions which will boost sustainable performance by adopting AI technology. Managers should launch initial projects which focus on essential high-impact zones to create visible advantages and gain company-wide trust. Organizations need to prioritize employee training that specifically prepares their workers to use AI tools effectively for their organizational goals. Managers need to create an environment of learning through teams that share knowledge and test AI applications per the findings of [
30]. Established processes for monitoring technological preparedness through evaluation of both system infrastructure and operational processes will reveal gaps needed for resolution. SMEs in developing economies can access customized AI solutions through partnerships between technology providers and research institutions as part of their collaborative ecosystem development. The combination of these implementation methods makes AI implementation result in meaningful achievements of sustainable performance goals.
5. Limitations and Future Research Recommendations
While this study makes important contributions, there are some notable limitations that should be discussed. First, the study was performed in Pakistan on SMEs, which can be a limiting factor for the results to be generalized across other geographical regions or larger enterprises. Though using a relevant testing method, the results may not be applicable to the broader population, as the unique socio-economic and cultural characteristics of Pakistan’s society and regulatory environment may have diverse implications for adoption of AI and technological readiness compared to other countries. Second, this was a cross-sectional survey study, collecting data at one specific time point. Although this approach helps analyze relationships among the variables, it does not reflect the temporal dynamics of tech adoption, workforce capacity building, and organizational performance. It is possible that a longitudinal approach could have generated richer insights into how these relationships evolved. Finally, this research examines only AI adoption and technology readiness, which represent only two aspects of a larger technology ecosystem. Other technologies such as IoT, blockchain, and big data analytics were not covered, possibly narrowing the study’s findings on technology melding.
With these limitations established, future studies can build in a range of ways. First, researchers should examine the generalizability of the findings by replicating studies in different geographical and industrial contexts. Studies comparing different regions or sectors could provide a wider view of how cultural and organizational factors impact the connection between technological integration, employee capacity building, and sustainable performance. Second, further research can be extended using longitudinal research design to understand the temporal changes associated with AI adoption, technological readiness, employee capacity building, and organizational performance. This would enable them to study the interaction of these variables and follow their evolution over time leading to deeper insights. Lastly, future research can further broaden the scope to include other emerging technologies such as big data analytics, blockchain, or IoT. It would also offer a more well-rounded perspective on the integration of technology and its implications for overall organizational performance and employee development.