How to promote employee engagement? A comparison of top-down versus bottom-up approaches

Employee engagement has attracted significant attention among academics and practitioners (Bailey, 2022), yet many organizations continue to struggle with how best to elevate and sustain employee engagement (Lee et al., 2023). Numerous studies have examined practices that appear to strengthen employee engagement (Alam et al., 2024; Chandni and Rahman, 2020), yet competing claims persist regarding the most effective means to achieve enduring outcomes (Byrne, 2022; Davis and Van der Heijden, 2024; Wittenberg et al., 2024).

On the one hand, researchers emphasize top-down approaches—commonly operationalized through (traditional) human resource management (HRM) practices—as effective pathways to stimulate employee engagement (Goyal et al., 2024a). These approaches include ability-enhancing HRM practices aimed at improving employees’ competencies and skills through training and development programs; motivation-enhancing HRM practices designed to encourage higher effort and proactive behaviors through performance-based compensation and rewards; and opportunity-enhancing HRM practices intended to empower employees through increased autonomy and involvement in decision-making (Li et al., 2022; Salvador-Gómez et al., 2023), which, in turn, support their performance (Shahzad et al., 2025). Despite their promise, these HRM practices may not always produce the immediate improvements in employee engagement that organizations often require (Johar et al., 2024), thereby igniting the need for further research to clarify and improve their effectiveness (1st research motivation).

On the other hand, growing evidence indicates that engagement can also emerge from bottom-up actions that offer employees greater autonomy and personal involvement in their jobs (Björk et al., 2021). Job crafting, for instance, empowers employees to engage in self-initiated changes at work (e.g. taking on additional tasks that match their interests, reorganizing the way tasks are performed, or proactively seeking out collaboration with colleagues) to achieve better alignment between their personal needs and preferences and organizational objectives (Roczniewska et al., 2023). This bottom-up perspective suggests that employees who proactively reshape their responsibilities can drive their own engagement more swiftly than those who await top-down directives (Tims et al., 2022), which, in turn, supports innovation performance (Rafiq et al., 2023). Yet, despite its apparent advantages, the effectiveness of job crafting relative to the (traditional) HRM practices above—which differs from HRM systems specifically designed to support job crafting (Hu et al., 2022)—remains relatively underexplored in promoting employee engagement (2nd research motivation).

Extant research on psychological ownership and organizational citizenship behavior (OCB) introduces additional layers to this debate. Psychological ownership, whereby individuals feel a sense of personal possession toward certain aspects of their work or workplace, is associated with stronger emotional attachment and sustained engagement (Dai et al., 2021; Zhang et al., 2021), which, in turn, contributes to job performance (Bai et al., 2024) and organizational imperatives (Jia et al., 2024). Similarly, scholars note that OCB—defined as voluntary actions beyond formal job duties (Organ, 1988)—positively enhances organizational performance and influences responsiveness to market pressures (Showkat et al., 2024; Vargas-Hernandez and González-Ávila, 2025). Although research indicates that both psychological ownership and OCB play a key role in supportive HRM practices (Goyal et al., 2024b), researchers caution that these factors alone may not sufficiently ignite engagement unless employees also have opportunities to modify their roles actively (Goyal et al., 2024a). Therefore, this study argues that deeper investigation is needed into how organizations can concurrently leverage top-down and bottom-up approaches alongside psychological ownership and OCB to sustain lasting employee engagement (3rd research motivation).

To critically advance this debate and integrate the identified research motivations, this article draws on social exchange theory (Blau, 1964) as its focal theoretical lens to explain the reciprocal interactions between employees and organizations (Cropanzano and Mitchell, 2005). Unlike other theories that are only tangentially applicable—such as stakeholder theory (Freeman, 1984), which emphasizes broader stakeholder interests rather than internal organizational exchanges, or self-determination theory (Ryan and Deci, 2000), which highlights employees’ autonomy, competence, and relatedness without fully capturing organizational–employee reciprocity—social exchange theory captures how organizational practices (top-down HRM) and opportunities provided by organizations for individual initiatives (bottom-up job crafting) serve as resources that employees reciprocate by demonstrating psychological ownership, OCB, and engagement. The research objective of this article is, therefore, to compare these two approaches—top-down HRM practices and bottom-up job crafting—to critically address three key research questions (RQs): whether one approach yields more immediate and substantial engagement benefits (RQ₁), whether psychological ownership and OCB explain how the effects of these approaches translate into employee engagement (RQ₂), and whether integrating both approaches represents the most effective strategy for enhancing employee engagement (RQ₃). Answering these RQs directly addresses ongoing tensions between traditional HRM and job crafting (1st research contribution), clarifies the unique and combined roles each plays in promoting employee engagement (2nd research contribution), and provides actionable insights for organizations seeking to elevate and sustain employee engagement (3rd research contribution).

HRM spans a broad set of interlinked practices—such as hiring, training, and talent retention—that aim to shape employee attitudes and behavior (Azam, 2023). Prior studies have shown that well-designed HRM practices can strengthen employees’ autonomous motivation (Han et al., 2024), reduce stress through work-life balance (Rashmi and Kataria, 2022), and enhance psychological resources that are crucial for employee engagement (Jose et al., 2024). However, recent research points out that these advantages may not always materialize (Johar et al., 2024). Some HRM strategies yield modest gains in engagement that dissipate when work environments are unpredictable, whereas others appear more enduring (Ererdi et al., 2022; Oliveira and Rocha, 2017). Critics thus question whether conventional HRM practices alone can deliver the short-term boosts in engagement organizations seek, especially in industries undergoing rapid change (Yalenios and d’Armagnac, 2023). From a social exchange theory perspective, employees may reciprocate supportive HRM with heightened engagement, but the degree of reciprocity can vary depending on individual and contextual factors (Presbitero, 2017; Saks, 2022; Urbini et al., 2021). This study contends that despite unresolved questions regarding the immediacy of HRM’s impact, recent evidence suggests that HRM practices, which by nature are ingrained in organizational routines and structures, generally provide essential and ongoing support for employee engagement (Pimenta et al., 2024). Accordingly, the following hypothesis is advanced:

A growing body of work argues that organizational practices alone may not suffice to foster the deeper sense of ownership and interest required for sustained engagement to perform at work (Ababneh, 2021; Bakker et al., 2014). Instead, employees themselves can initiate changes in their job responsibilities, social interactions, or work methods—a process termed “job crafting” (Roczniewska et al., 2023; Tims et al., 2022). This self-directed reshaping is believed to align personal preferences with organizational objectives more effectively than standard top-down interventions (Björk et al., 2021; Tims et al., 2022). Yet, empirical evidence on job crafting’s effectiveness in driving engagement remains inconsistent and limited. Some studies indicate that employees who craft their tasks report heightened engagement (Bakker et al., 2014) and well-being (Ghazzawi et al., 2021), whereas others question whether employees lacking autonomy or self-efficacy would be equally motivated or able to craft their jobs (Miraglia et al., 2017). Divergent findings also raise questions regarding job crafting’s universality and staying power, as pursuing more challenging job demands can unintentionally increase workload, thus consuming additional resources and eventually undermining employees’ capacity to remain engaged over time (Birman et al., 2024). Despite these debates, social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005) suggests that employees are likely to reciprocate autonomy and organizational support with higher engagement, leading many scholars to conclude that job crafting offers a potentially faster route to engagement by giving employees direct control over their daily experiences (Roczniewska et al., 2023; Tims et al., 2022). As such, the following hypothesis is established:

Although HRM is often praised for cultivating a supportive work environment (Azam, 2023), the mechanisms that translate HRM practices into stronger engagement remain contested. Scholars argue that psychological ownership—where employees feel a personal sense of possession toward their work—is essential for deep commitment and devotion (Van Dyne and Pierce, 2004). Indeed, prior research has shown that practices focusing on fair compensation (Kim, 2025) and participative decision-making (Verkuyten, 2025) can spark this sense of ownership. Yet, not all studies confirm that HRM automatically drives psychological ownership, with some noting that employees may dismiss HRM practices such as empowerment as transactional gestures if they perceive a lack of genuine inclusion or autonomy (Katou, 2022). Tensions thus persist regarding when HRM triggers psychological ownership strongly enough to elevate engagement. Indeed, scholars have offered mixed views: some emphasize the positive potential of well-aligned HRM systems (Baykal and Bayraktar, 2022), whereas others note that surface-level practices fail if they do not address underlying employee needs (Boonsiritomachai et al., 2022). In line with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), this study posits that HRM can effectively facilitate psychological ownership—provided employees perceive HRM practices as genuinely supportive, valuable, and responsive to their needs (e.g. ability-, motivation-, and opportunity-enhancing practices; Li et al., 2022; Salvador-Gómez et al., 2023)—and that psychological ownership mediates the relationship between HRM and employee engagement because such ownership creates deeper emotional bonds and personal accountability, driving employees to reciprocate organizational support through increased engagement (Baykal and Bayraktar, 2022; Van Dyne and Pierce, 2004). Consequently, the following hypothesis is formulated:

Job crafting, unlike standardized HRM practices, grants employees a distinct sense of control and identity in shaping their work experiences (Roczniewska et al., 2023; Tims et al., 2022). This autonomy fosters psychological ownership by satisfying deeper motives such as self-efficacy and belonging (Jing and Yan, 2022). Indeed, prior research highlights that employees who actively reshape their tasks (Shin et al., 2018; Tims et al., 2015) and social interactions (Doden et al., 2024) often develop stronger personal connections to their roles, significantly enhancing their engagement. Critics caution, however, that not all forms of job crafting ensure meaningful ownership. An overemphasis on minor task adjustments, for example, may fail to generate the deeper sense of meaning and attachment necessary for lasting commitment (Letona-Ibañez et al., 2021). Moreover, the effectiveness of job crafting can vary according to the availability of job resources (Tims et al., 2022) and individual employee readiness (Szőts-Kováts and Kiss, 2023). In line with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), this study contends that when organizations genuinely empower employees to craft their roles—thus providing valuable resources—employees reciprocate through increased psychological ownership, which mediates the relationship between job crafting and employee engagement because employees who experience ownership feel a deeper emotional investment and personal responsibility toward their work, prompting sustained engagement (Jing and Yan, 2022; Tims et al., 2022). Hence, the following hypothesis is posited:

A parallel debate centers on whether HRM practices produce stronger engagement by encouraging employees to go beyond prescribed duties, commonly referred to as OCB (Organ, 1988). Scholars have documented a link between a positive work environment and OCB, suggesting that employees reciprocate supportive HRM by helping co-workers and upholding a culture of cooperation (Carter et al., 2021). Nonetheless, counterarguments exist, whereby some scholars maintain that employees may perform extra-role behaviors only under certain conditions, such as fair reward systems (Mach et al., 2025) and genuine opportunities for advancement (Shang et al., 2021). Moreover, while research has shown that OCB can strengthen employees’ commitment and intent to stay (Halid et al., 2024), skeptics caution that OCB alone may not suffice to boost engagement unless employees perceive a clear two-way exchange between themselves and the organization (Bolino et al., 2024). In line with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), this study argues that employees engage in OCB—and subsequently become more deeply engaged—when they interpret HRM practices as genuine organizational support worthy of reciprocal effort because such perceptions of reciprocity foster a sense of emotional attachment and obligation, motivating employees to consistently perform beyond prescribed roles, thereby strengthening their overall engagement (Bolino et al., 2024; Carter et al., 2021). Therefore, the following hypothesis is proposed:

Job crafting often enables employees to structure their tasks and social interactions in ways that align with personal strengths and interests (Roczniewska et al., 2023; Tims et al., 2022). This freedom can stimulate a proactive attitude toward colleagues and the organization (Srivastava and Pathak, 2020). Yet, the relationship between job crafting and OCB is not always straightforward. For instance, employees who encounter strict policies or minimal autonomy may be unable to craft their roles adequately and thus refrain from discretionary behaviors; however, when job crafting does occur, employees may feel motivated to support co-workers, handle shared challenges, and invest in tasks beyond formal role requirements (Magdaleno et al., 2023). In line with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), this study contends that when organizations grant employees sufficient autonomy and opportunities to craft their roles, employees reciprocate by engaging in higher levels of OCB, which, in turn, enhances employee engagement because discretionary behaviors—such as taking initiative beyond formal duties and voluntarily supporting colleagues—strengthen employees’ social ties and personal investment in their work (Magdaleno et al., 2023; Srivastava and Pathak, 2020). Thus, the following hypothesis is put forth:

The service sector depends heavily on human interaction and interpersonal skills, clearly distinguishing it from the manufacturing sector, which mainly relies on machinery for production. This distinction compels service-driven organizations to emphasize employee engagement as a strategic means of sustaining competitive advantage (Bhasin et al., 2019). Within the service sector, the Indian information technology (IT) services industry represents an especially relevant context because it is highly dynamic, continually shaped by technological advances, and marked by intense competition for skilled personnel (Gupta and Basole, 2020). Noteworthily, India’s global prominence as one of the largest IT outsourcing providers (Statista, 2025) underscores the industry’s reliance on employee engagement, as employees are integral to delivering consistent service quality, innovative solutions, and client satisfaction (Mahendramohan et al., 2023). Recognizing these strategic demands, this study positions the Indian IT services industry as a particularly suitable context for investigating employee engagement—an area where maintaining consistently high employee involvement at work is challenging yet crucial (Hasan et al., 2021). Rapid growth and continuous change within IT services further necessitate strategic measures that effectively balance structured, top-down HRM interventions with flexible, bottom-up initiatives such as job crafting. Such balanced strategies ensure employees maintain sufficient motivation and resilience to address evolving client expectations and competitive pressures. Unlike manufacturing sectors, where automation frequently substitutes for human effort, IT services require specialized human capital capable of integrating technical knowledge with advanced relational competencies. In this regard, the selected industry aligns closely with both the theoretical arguments and practical objectives of this study. In engaging with this specific context, the present study examines whether competing claims about the effectiveness of top-down (HRM) versus bottom-up (job crafting) approaches to employee engagement can be reconciled and explores how psychological ownership and OCB mediate these relationships, recognizing that sustained employee engagement is paramount in an industry dependent upon the ongoing engagement of its workforce.

A survey was conducted using a questionnaire with items adapted from past studies. All items were answered on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). First, this study adapted ability-motivation-opportunity (AMO) based HRM practices such as those that are ability-enhancing (e.g. competency hiring, training), motivation-enhancing (e.g. financial and non-financial reward), and opportunity-enhancing (e.g. autonomy, career development, open participation) (Mehralian et al., 2021; Nadeem and Rahat, 2021). For instance, HRM1 is the selection of new hires based on their competency and skills, HRM2 provides continuous training opportunities to employees (ability-enhancing practices), HRM3 offers rewards comparable to other organizations for the same job and profile (motivation-enhancing practices), HRM4 includes career development opportunities within the IT company, and HRM5 involves employee participation in decision-making (opportunity-enhancing practices). Second, job crafting measures were adapted from Leena et al. (2009), utilizing individual and collaborative job crafting—namely, “On your own, change the way you do your work to make it easier for yourself” (JC1) and “Decide together with your co-workers to organize special events at your workplace” (JC2). Third, psychological ownership measures were adapted from Van Dyne and Pierce (2004), including “This is my company” (PO1), “I sense that I own this company” (PO2), and “I feel a very high degree of personal ownership” (PO3). Fourth, OCB measures were adapted from Lee and Allen (2002), including “Assisted others with their duties” (OCB1), “Willingly gave time to help others who had work-related problems” (OCB2), and “Expressed loyalty towards the company” (OCB3). Fifth, employee engagement measures were adapted from Schaufeli et al. (2006), measuring vigor – “At my work, I feel bursting with energy” (EE1); dedication – “My job inspires me” (EE2); and absorption – “When I am working, I forget everything else around me” (EE3).

A combination of a random and purposive sampling approach was adopted to administer the survey and collect the data. The random sampling was applied to select service providers in the IT industry while the purposive sampling approach was used to select respondents who are operating at the executive-level with service responsibilities in this industry. In particular, 500 surveys were administered to executive-level workers in the IT services industry in Northern India, of with 364 were returned and usable (±73% usable response rate). Of the 364 respondents, 54% were male and the remaining female. The majority of participants were graduates (63%) and 27% were postgraduates. Sixty percent of respondents belonged to the 20–30 age group with five or fewer years of experience. Thirty-five percent were in the 30–40 age group with up to 10 years of experience, and the rest were above 40 years old with extensive experience.

This study incorporates both structural equation modeling (SEM) and artificial neural network (ANN) analyses in a single investigation to enable a richer understanding of the factors driving employee engagement (Shahzad et al., 2020). SEM, which is hypothesis-driven, is complemented by the data-driven capabilities of ANN, thereby enabling the study to rigorously test its theoretical assumptions while also uncovering complex patterns that may further explain employee engagement. This dual deductive-inductive approach, therefore, reinforces the rigor of the study’s conclusions about employee engagement.

The study could potentially suffer from the problem of common method bias (CMB) since the respondents answered the questionnaire for all variables at the same time (Podsakoff et al., 2003). To address this issue, a single-factor test and a latent marker variable test were conducted (Lim, 2025). The single factor test revealed a variance (43.38%) below the 50% maximum threshold for significant CMB, whereas the latent marker variable test showed no significant differences across two models—one incorporating the marker variable and the other without it (Table 1). Therefore, CMB is not a significant concern for this study.

Table 2 presents the statistical results required to evaluate the reliability and validity of the measurement model. In terms of reliability, the Cronbach’s alpha, composite reliability (rho_c), and Dijkastra-Henseler’s rho (rho_A) values for all variables are above the minimum threshold of 0.70, which indicates that the measures for the variables are reliable (Hair et al., 2017). In terms of convergent validity, the average variance extracted (AVE) for all variables is above the minimum threshold of 0.50, which means that the measures are closely related to their respective variables (Fornell and Larcker, 1981). In terms of discriminant validity, the square root of the AVEs are larger than the correlations between variables (Fornell and Larcker, 1981) and the heterotrait-monotrait (HTMT) ratio of correlations are less than the maximum threshold of 0.85 (Henseler et al., 2015), which suggests that the variables are distinct and do not overlap significantly. Therefore, the measurement model can be said to possess adequate reliability as well as convergent and discriminant validity.

Partial least squares SEM (PLS-SEM) was employed to test the hypotheses. A nonparametric bootstrap with 5,000 samples was used to assess the significance of coefficients in PLS-SEM (Sarstedt et al., 2022), testing the impact of HRM and job crafting on employee engagement through psychological ownership and OCB. The R² for psychological ownership, OCB, and employee engagement were 0.529, 0.444, and 0.540, respectively, indicating a high level of explanatory power.

Figure 1 and Table 3 Panel A illustrate the main effects, showing an insignificant direct effect of HRM (β = 0.011, p = 0.834 > 0.05, t = 0.210 < 1.96, f² = 0.000) and a significant direct effect of job crafting (β = 0.213, p =0.006 < 0.01, t = 2.727 > 2.576, f² = 0.037) on employee engagement, with the strength of the latter relationship characterized by a small effect size (Cohen, 1988; Lim, 2025). Thus, H₂ is supported, but not H₁.

Table 3 Panel B reports the mediating effects. The indirect effect of HRM to employee engagement through psychological ownership was significant (β = 0.071, p = 0.009 < 0.01, t = 2.632 > 2.576), as was the path from job crafting to employee engagement via psychological ownership (β = 0.143, p = 0.000 < 0.01, t = 4.067 > 2.576), implying that psychological ownership facilitates this indirect effect through full and partial mediation, respectively. Therefore, H_3a and H_3b are supported. Similarly, the indirect effect of HRM to employee engagement through OCB was significant (β = 0.079, p = 0.001 < 0.01, t = 3.425 > 2.576), as was the path from job crafting to employee engagement via OCB (β = 0.188, p = 0.000 < 0.01, t = 4.476 > 2.576), implying that OCB facilitates this indirect effect through full and partial mediation, respectively. Thus, H_4a and H_4b are supported.

ANNs are increasingly acknowledged in business research as a powerful tool for simulating human decision-making and judgment processes (Haykin, 2004). They are distinct from traditional regression models as they do not necessitate linearity assumptions (Wilson and Bettis-Outland, 2020), and thus, they can serve as a suitable data-driven complement to the hypothesis-driven SEM (Shahzad et al., 2020).

In our study, we employed the feed-forward back-propagation (FFBP) multilayer perceptron (MLP) technique to train the data and assign relative weights to the predictors. To mitigate the risk of overfitting, we used a 10-fold cross-validation method, allocating 90% of the data for training and 10% for testing. Training data constitutes a subset of the original data dedicated to training the model, whereas testing data is used to verify the model’s accuracy.

The basic structure of the neural network, as depicted in Figure 2, consists of three primary layers: input, hidden, and output. The input layer houses the predictors (HRM, job crafting, psychological ownership, and OCB). The hidden layer contains unseen units or nodes, the value of each can be characterized as a function of the predictive variables. The specific function shape is determined by the network type and user-defined parameters. The responses are housed in the final layer, namely, the output layer, which in this context corresponds to employee engagement.

The network accommodates bias in a manner similar to the inclusion of an intercept in a linear equation. Alongside the weighted total of the inputs to the neuron, bias acts as an additional parameter in neural networks, altering the output and effectively shifting the activation function to the right or left. This is depicted in Figure 2, where dark lines indicate a strong relationship between inputs (predictors) and output (outcomes).

We designed the input and hidden layers using MLP and sigmoid activation functions, implemented in the Statistical Package for Social Science (SPSS). The prediction accuracy was assessed via the root mean square error (RMSE) values (Table 4). The average RMSE values for all ANN models ranged from 0.055 to 0.159 for training data and 0.012 to 0.152 for testing data. The minimal differences between training and testing RMSE values, combined with their overall low magnitude—especially in the testing set—indicate that the ANN models reliably capture the relationships between the predictors and the outcome.

To evaluate the normalized importance of each input neuron, we also conducted a sensitivity analysis. We first computed the average relative importance for all 10 neural networks. We then normalized these values by dividing each by the highest importance value, converting the result into a percentage. As illustrated in Table 5, the normalized importance varied from 44.5% to 100%. Among the factors impacting employee engagement, OCB was identified as the most significant, followed by job crafting, psychological ownership, and HRM.

This study critically examines the relative effectiveness of top-down (HRM practices) versus bottom-up (job crafting) approaches in promoting employee engagement, addressing ongoing debates about the potential of these strategies. Drawing on social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), this study specifically investigates whether job crafting provides more immediate and substantial benefits to employee engagement compared to HRM practices (RQ₁), how psychological ownership and organizational citizenship behavior serve as mediating mechanisms translating these approaches into employee engagement (RQ₂), and whether integrating both approaches represents the most effective strategy for achieving lasting employee engagement (RQ₃). The subsequent paragraphs elaborate upon the key empirical findings derived from these RQs, outline their theoretical implications, and present actionable managerial insights to guide organizations seeking to effectively elevate and sustain employee engagement.

To begin, the first key finding of this study indicates that HRM does not exert a significant direct effect on employee engagement, refuting H₁. This finding diverges from prior research suggesting that HRM practices positively drive engagement (Aktar and Pangil, 2018; Sivapragasam and Raya, 2018). A possible reason for the divergence is that HRM practices alone may be perceived by employees as transactional rather than relational, especially if they lack genuine opportunities for personal initiative and involvement, thereby limiting their perceived value as resources employees would feel obligated to reciprocate, consistent with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005). Accordingly, the first theoretical contribution is that this outcome challenges the widely held assumption that HRM always enhances engagement by highlighting that HRM effectiveness in promoting engagement critically depends on employees’ subjective appraisal of HRM practices as meaningful and genuinely supportive (Katou, 2022; Presbitero, 2017). This insight advances ongoing debates regarding the conditions under which HRM interventions fail to secure the positive behaviors they are designed to elicit.

Contrastingly, the second key finding confirms that job crafting has a clear, positive relationship with employee engagement, supporting H2. This observation aligns with earlier work illustrating how employees who personalize their roles tend to experience deeper investment in their day-to-day tasks (Bakker, 2010; Chen et al., 2014). The result also addresses debates on whether bottom-up behaviors provide a more rapid path to engagement compared to conventional, top-down efforts because, as suggested by social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), employees view opportunities to craft their roles as valuable resources directly provided by their organizations, thus motivating them to reciprocate immediately through increased engagement. Contrasted with the first major finding—where HRM alone failed to directly influence engagement—this clearly answers RQ₁ by highlighting that job crafting, as a bottom-up approach, delivers more immediate and substantial engagement benefits than top-down HRM practices. Building on this, the second theoretical contribution is that these findings expand social exchange theory by illustrating how the perception of organizational support through autonomy and personalized control over work roles directly fosters reciprocal employee engagement.

Building on the above, the third key finding demonstrates that HRM can still promote engagement indirectly by fostering psychological ownership, supporting H3a. This finding indicates that HRM practices, when perceived by employees as genuine signals of organizational support rather than transactional gestures, cultivate a sense of personal investment and ownership in their work roles. In line with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), employees who interpret HRM interventions as meaningful investments by their organization feel obligated to reciprocate positively, and psychological ownership provides the psychological mechanism for this reciprocal relationship to occur (Van Dyne and Pierce, 2004). Employees thus respond more deeply and sustainably because they perceive a genuine exchange of support. Consequently, the third theoretical contribution is that these findings specify a condition under which HRM fulfills its potential: HRM-based initiatives must cultivate authentic feelings of ownership rather than mere compliance to generate beneficial attitudinal outcomes. This clarifies why some organizations observe significant gains from HRM while others experience minimal effects, refining social exchange theory by demonstrating that psychological ownership specifically functions as a decisive mediator in facilitating genuine employee reciprocity.

Complementing the above, the fourth key finding highlights that job crafting also shapes engagement through psychological ownership, supporting H3b. This result occurs because employees interpret the autonomy to personalize their roles as valuable organizational support, consistent with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005). Specifically, when organizations empower employees to craft their tasks, employees perceive this empowerment as a meaningful investment in their autonomy and well-being, prompting them to reciprocate through heightened psychological ownership. Such ownership intensifies employees’ personal stakes in their roles, motivating sustained engagement, thereby aligning closely with Wang et al.’s (2018) call to explore the underlying motivational mechanisms of job crafting. Drawing on this, the fourth theoretical contribution is that this result clarifies precisely why job crafting enhances engagement: feelings of personal possession—fueled by enhanced self-efficacy and deeper belonging—form the psychological conduit through which employees reciprocate organizational investments with stronger, sustained engagement. This refinement enriches the literature on proactive work behavior by explicitly identifying psychological ownership as a key mediator in the job crafting–engagement relationship.

Extending further, the fifth key finding indicates that OCB fully mediates the relationship between HRM and engagement, supporting H4a. This finding can be explained through social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), which suggests that supportive HR practices, when perceived by employees as genuine acts of organizational goodwill, motivate employees to reciprocate by willingly performing extra-role behaviors beneficial to colleagues and the organization (Halid et al., 2024). Employees engaging in these discretionary behaviors develop stronger social ties, experience a deeper sense of belonging, and perceive a more meaningful connection to their workplace, all of which subsequently enhance their engagement. Extending this logic, the fifth theoretical contribution is that OCB emerges as a pivotal mechanism translating HRM inputs into engagement, clarifying that engagement rises specifically when employees interpret their discretionary efforts as a fair reciprocal exchange in response to meaningful organizational support. This mediator enriches prior knowledge by demonstrating that structural HR interventions alone are insufficient to stimulate engagement unless they also encourage voluntary extra-role behaviors based on perceived reciprocity.

Further reinforcing this logic, the sixth key finding reveals that OCB partially mediates the link between job crafting and engagement, supporting H4b. This finding aligns with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), suggesting that when employees experience autonomy to reshape their roles, they perceive this autonomy as meaningful organizational support that warrants reciprocation. Employees thus respond by willingly engaging in discretionary behaviors that benefit their colleagues and organization (Magdaleno et al., 2023; Srivastava and Pathak, 2020). These cooperative and supportive actions subsequently create stronger interpersonal bonds and deeper feelings of organizational identification, intensifying their overall engagement. Further clarifying, the sixth theoretical contribution is that this result strengthens arguments that bottom-up initiatives such as job crafting not only yield personal benefits but also generate social resources, through OCB, that reinforce engagement at a collective level. Through explicitly demonstrating how autonomy-driven role modifications foster reciprocal extra-role behaviors, this contribution clarifies a critical mechanism by which bottom-up approaches enhance employee engagement beyond the capabilities of the traditional, top-down counterparts.

Finally, the seventh key finding stems from the ANN, which identifies OCB as the most influential predictor of employee engagement, followed by job crafting, psychological ownership, and HRM. This result emerges from a sensitivity analysis of the average relative importance of each input neuron, confirming that OCB surpasses other variables in shaping employee engagement. This result aligns closely with social exchange theory (Blau, 1964; Cropanzano and Mitchell, 2005), as it indicates employees’ discretionary behaviors strongly reflect perceived reciprocity—when employees willingly perform extra-role actions, they signal a deeper relational connection to their organization, producing a stronger engagement response than formally mandated or routine behaviors. Grounded in this reasoning, the seventh theoretical contribution is that the ANN approach, free from linear assumptions, highlights the comparative strength of OCB as a predictor and thus provides empirical validation for the theoretical position that discretionary, voluntary behaviors represent especially potent forms of reciprocation in organizational exchanges. This insight supports the contention that extra-role behaviors—particularly those encouraged by autonomy-driven job crafting—serve as critical drivers of employee engagement. This insight also encourages researchers and practitioners to incorporate data-driven approaches alongside traditional hypothesis-based models, reinforcing the social exchange theory by demonstrating that voluntary, discretionary actions generate the strongest reciprocal responses from employees.

Building on the theoretical insights outlined above, the first managerial implication drawn from these findings is that organizations seeking immediate and tangible improvements in employee engagement should prioritize job crafting initiatives. Unlike HRM practices, which showed no significant direct impact on employee engagement, job crafting directly influences engagement by empowering employees to personalize their tasks and interactions. Practically, managers can achieve this by explicitly providing employees with autonomy to reshape their roles—for instance, by encouraging employees to take ownership of tasks aligned closely with their personal strengths and interests or by allowing flexible task arrangements and self-managed teams. Clearly communicating to employees that such bottom-up initiatives are welcomed and valued will help foster immediate increases in engagement by enhancing their sense of autonomy and intrinsic motivation.

However, the second managerial implication emphasizes caution against dismissing HRM practices entirely, as this study has highlighted indirect but meaningful pathways through which HRM can still positively affect employee engagement. Managers should actively leverage HRM interventions by ensuring these practices explicitly nurture psychological ownership and OCB. For example, managers can strengthen psychological ownership by involving employees in meaningful decision-making processes, clearly articulating how employees’ roles contribute to organizational success, and openly acknowledging their contributions. Similarly, fostering OCB can be accomplished by establishing fair reward systems that recognize and reward extra-role behaviors, publicly celebrating teamwork and collaborative achievements, and creating structured opportunities for employees to assist colleagues beyond their formal duties. Such actionable steps ensure HRM practices become genuinely supportive rather than superficial, thus translating indirectly into sustained employee engagement.

Lastly, the third managerial implication is that organizations seeking to maximize employee engagement in the long run should adopt a hybrid strategy combining both bottom-up and top-down approaches. Given that job crafting yields immediate benefits through autonomy and empowerment, managers might first introduce targeted job crafting opportunities for more experienced and skilled employees who are ready to self-direct their roles. Concurrently, structured HRM practices remain essential, particularly for employees who may not yet possess the confidence, readiness, or skills required for autonomous role modification. Managers can implement phased programs—initially emphasizing autonomy-driven job crafting initiatives among leaders or senior personnel while progressively rolling out supportive HRM systems to equip other employees with the resources, skills, and confidence needed for future empowerment. This phased and complementary approach ensures sustainable engagement by simultaneously cultivating autonomy, psychological ownership, and OCB across the entire organization.

This study critically compared the effectiveness of top-down HRM practices and bottom-up job crafting approaches in promoting employee engagement within the service industry. Given the people-intensive nature of services, ensuring high levels of employee engagement is imperative for organizational performance and sustained competitive advantage. Through an integrated analysis employing both SEM and ANN, this study demonstrated that job crafting directly and immediately enhances employee engagement, whereas HRM practices contribute indirectly through psychological ownership and OCB. Consequently, organizations looking for immediate improvements should prioritize empowering employees to proactively reshape their roles. However, lasting and sustainable employee engagement requires a balanced approach—integrating autonomy-enhancing job crafting initiatives with structured HRM practices strategically designed to cultivate psychological ownership and encourage OCB. Thus, organizations must thoughtfully combine both approaches to fully realize the potential of employee engagement.

While the current study focused on positive employee behavior, future research could also address negative engagement like counterproductive workplace behavior and service failures, as well as the solutions to overcome them such as error management culture (Aggarwal et al., 2024). In addition to examining the employees who engage in OCB, the perceptions of those who receive OCB should be considered, including in a longitudinal context. A multi-level approach incorporating different tiers of employees and their superiors could provide a clearer understanding of the phenomenon. Furthermore, the generalizability of the findings is also limited. As mentioned, the importance of cross-validation on diverse employee groups highlights the need for retesting the model on larger samples and in various cultural contexts in future studies. Moreover, this study utilized self-reported, cross-sectional data in the Indian IT services industry, which constrain the causal inferences that can be made. To widen the scope of this study, comparative analyses between various industries and other countries could be conducted, whereas combining quantitative and qualitative methodologies (Takona, 2024) would offer a more in-depth understanding of how employees perceive their workplace environments across these industries and countries.