The cheap back office becomes a transformation driver: managing ambidexterity in shared services

Recent years have brought significant changes to service businesses with digital technologies reconfiguring offerings, delivery systems, and business models (Madlberger, 2023). As Kandampully and colleagues (2021, p. 219) argue, digitalization expands how firms “reach, inform, engage, sell to, learn about, and provide service to customers”, thereby catalyzing service transformation at both firm and subunit levels. This study investigates the service transformation of shared service organizations (SSOs) of multinational corporations.

Since the 1990s, SSOs have centralized transactional activities (e.g. finance, HR, IT), blending Taylorized process control with ERP-enabled workflows and location advantages to deliver efficiency at a global scale (Gospel and Sako, 2010; Richter and Brühl, 2021). The Finance and HR SSO of SAP AG is a case in point. As is typical of SSOs, it consists of several shared service centers (SSCs) – offices where the work of SSOs for their internal customers in the company that has set up the SSO – here SAP–is performed. Three SSCs were set up in Argentina, Prague, and Singapore to serve all 100-plus SAP business units across the Americas, Europe, the Middle East, and Asia-Pacific regions as their internal customers. It delivers cost savings by standardizing and bundling tasks in Finance, Accounting, and HR, and providing them as an internal service provider from just three locations, showcasing SAP's consistent use of its own ERP software across the three locations for greater efficiency. Traditional transactional services moved into the SSO include payroll processing, travel expenses, accounts receivable, and accounts payable (Vollmer and Rasper, 2013).

Yet, digitalization poses a significant challenge for SSOs, primarily because it disrupts the traditional value proposition of these organizations. This inflection becomes a case of service transformation, where service organizations may adopt new pathways that reconfigure offerings, delivery systems, and value propositions (Kandampully et al., 2021). SSOs have provided value to companies through cost savings from economies of scope, efficiency gains, and labor cost arbitrage, but these benefits are eroding as digitalization advances (Sako, 2015). Robotic process automation (RPA) and other digital technologies have increasingly automated mainly the less complex, more repetitive aspects of professional work, but in some cases, also non-routine tasks (Frey and Osborne, 2017). This trend has also arrived in SSOs, where precisely these tasks, which are especially vulnerable to automation, have been moved in many companies (Kearney, 2016). In response, SSOs rebalanced their service portfolios and recalibrated their value propositions, shifting toward more complex, knowledge-intensive services and transformation-oriented activities (Klimkeit and Reihlen, 2022; Colombo and Beuren, 2023). According to a recent industry report, 43% of SSOs now provide data analytics/business analytics services, the same percentage offers tax support, and SSOs increasingly extend into supply chain management, risk management, advisory, and legal services (SSON, 2025). To date, however, there has been little inquiry into how digital technologies reshape professional service work inside SSOs. In particular, how SSOs transition from the old efficiency to an innovation regime is still largely unexplored.

While service transformation models map key drivers and approaches (Kandampully et al., 2021; Madlberger, 2023), they offer limited insights into how transformation is managed as it unfolds. Kandampully et al. (2021) emphasize that service transformation proceeds through multiple steps rather than a single shift and explicitly ask how these changes should be managed (p. 226). Relatedly, Zaki (2019, p. 431) highlights the central tension of fostering innovation and risk-taking while continuing to run the existing business. In response to these calls, we adopt an ambidexterity perspective as a lens for explaining service transformation.

Drawing on research on organizational ambidexterity (see for an overview Ruano-Arcos et al., 2024; Turner et al., 2013), we explore the evolving role of SSOs, particularly in managing the paradoxical challenges they face today. The ambidexterity lens captures the dynamic nature of SSOs' service transformation, highlighting how they adapt to a changing digitalized business environment by maintaining operational efficiency while exploring new knowledge-intensive avenues for growth. Paradoxically, the SSO, the traditional organizational locus of exploitation, optimized for efficient, routine execution, is emerging as a hub for exploration. During the transition, if not beyond, exploitation and exploration must therefore co-exist, making ambidexterity a design imperative.

This raises our guiding question: How does the transformation of SSOs as internal service providers unfold as they transition from traditional exploitation to innovation-driven exploration, and which ambidextrous strategies, organizational dynamics, and capabilities enable this shift? Empirically, we conducted 38 interviews with 41 managers across nine SSCs in seven SSOs in the Czech Republic, Singapore, Malaysia, India, and the Philippines, as well as two interviews with two partners and one senior manager of an advisory firm that advises on shared services.

This study makes the following two contributions: First, based on our analyses, we provide a granular insight into the service transformation of SSOs by showing two modes of ambidexterity, which we refer to as the “transformation project mode” and “the moving up the value chain mode”. These two distinctive modes shape the transition from a predictable and efficient routine operation to a new organizational configuration that embraces more innovative, knowledge-intensive activities. Second, our findings also have implications for leading service workers. Traditionally, the work setting of SSOs has often resulted in “demeaning experiences” (Lepistö et al., 2018, p. 508) for service workers. Yet, with the emerging role of SSOs, we outline how SSOs' talent management, professional work, and management are becoming more like those in traditional professional settings, such as professional service firms.

We conceptualize service transformation as an organization-spanning, cyclical reconfiguration of value creation logics, structures, and offerings (Kandampully et al., 2021). Service transformation suggests a systemic reorientation in which the firm's strategy, structure, capabilities, and culture are reconfigured to explore new business opportunities (Hitt et al., 2011) or to respond to pressures from disruptive innovation that incumbents overlook due to value-network and resource-dependence constraints (Christensen, 2015). This reorientation is visible, for instance, in the way hospitality and tourism services are being redesigned around smart, data-rich environments that enable new forms of value co-creation (Buhalis et al., 2019), and in firms' moves from isolated smart services toward platform-based value creation that requires new IT, governance, and relational capabilities (Beverungen et al., 2021).

Service transformation arises from configurations of mutually reinforcing conditions. Internally, resources, structures, and culture set the feasibility and pace of change, while leaders adopting a revolutionary stance treat strategies, processes, and even organizational identity as adjustable to serve a higher-order service vision (Kandampully et al., 2021). Externally, digital technologies (e.g. AI, IoT, cloud computing) operate as principal enablers of reconfiguration (Madlberger, 2023). Field shocks and societal well-being demands, most visibly COVID-19, can translate environmental pressure into both transient adaptations and enduring innovations (Madlberger, 2023).

Following Kandampully et al. (2021), service transformation unfolds along three pathways. Reformation denotes incremental adaptation: firms wait for technologies to mature, then digitize existing processes and marginally refine offerings, retaining the prevailing logic while risking inertia. Renovation orchestrates selective destruction alongside renewal: leaders optimize existing routines to release resources, develop new capabilities, and progressively reconfigure the service system. Reincarnation is metamorphic: when the incumbent value logic becomes untenable, leaders initiate cycles of creative destruction, recast organizational identity, and migrate to a new value architecture. Taken together, these pathways clarify how firms sequence transformation in response to changing contingencies.

The First Wave: Globalization and Industrialization of SSOs: Since the 1990s, many companies have set up Shared Service Organizations (SSOs) in which large parts of corporate functions, such as accounting, HR, or IT for various business units, are bundled in semi-autonomous units. These SSOs carry out their work as internal service providers for many or all business units of the company, as the customers, under centralized governance (for a review see Richter and Brühl, 2017). In contrast to outsourcing activities, multinational corporations (MNCs) have retained many of these activities in-house to ensure control over confidential information and enhance the coordination of SSOs with business unit operations (Bondarouk, 2014). Following Bergeron (2003, p. 3), shared services is “a collaborative strategy in which a subset of existing business functions are concentrated into a new, semi-autonomous business unit … designed to promote efficiency, value generation, cost savings, and improve service for the internal customers of the parent corporation, like a business competing in the open market”.

The first wave's transformation logic aligns with the renovation logic (Kandampully et al., 2021): companies reframed their value proposition from local responsiveness to enterprise efficiency. Centralization promised scale economies, specialization, and standardization. Many shared services were rationalized through Tayloristic principles of controlling labor through Scientific Management, separating the conception and execution of work (Cooper and Taylor, 2000), to improve efficiency and control. Service workers are treated as replaceable “components,” and brainwork is removed from the shop floor by separating conception from execution. Following Howcroft and Richardson (2012, p. 111), standardization became a means to reframe skilled back-office service work as a routine job, resulting in, as Braverman (1974, p. 113) argues, a “dehumanization of the labor process”. Service work became “working in a repetitive, deskilled job, with relatively low pay and little prospects of promotion” (Cooper and Taylor, 2000, p. 574). As Mezihorak (2018, p. 840), in a case study of SSOs, concludes: “All these developments contribute to the increased fragmentation of labor processes into specific tasks (and sub-tasks) and their increasingly specific codification, which assures decreasing complexity. This consequently makes it easier to externalize, control, and finally, commodify the activities concerned”.

Commodification enabled a second lever: location arbitrage. Typical locations such as India, the Philippines, or Eastern Europe were favorable because of their financial attractiveness, the availability of skilled labor, and a favorable business environment (Kearney, 2016). The SSO of an MNC typically has one or more SSCs in each of the Americas, Europe, and Asia. Information and communication technology (ICT)- driven workflows using ERP systems align well with this managerial logic, as they further improve efficiency, control, and thus competitiveness of SSOs in a Neo-Tayloristic fashion.

The Second Wave: Digitalization and the Recomposition of Value: Currently, we witness a second wave where the automation of repetitive, transactional processes by technologies such as robotic process automation (RPA) is reconfiguring the shared-service model, eroding location-based, low-cost processing advantages (Pérez-Morón and Marrugo-Salas, 2021; Carpenter, 2024). It forces a strategic re-positioning toward more knowledge-intensive services (Chandok et al., 2016; Ferreira and Janssen, 2023). With the rise of digital technologies and service automation, the rules of the game for SSOs are being radically transformed or disrupted (Li et al., 2025). An increasing number of service activities are performed through digital agents that can be executed from any location worldwide (Kearney, 2016), potentially resulting in job losses in SSOs (Figueiredo and Pinto, 2021). Under these conditions, SSOs lose their traditional advantages, which are founded on blending standardization and control of routine work with the exploitation of locational advantages, especially labor costs. Unsurprisingly, SSO management is motivated to reinvent its strategy and position within the MNC, becoming a more strategic partner that contributes to the core business (Pinto et al., 2025). Klimkeit and Reihlen (2022) investigated the impact of digitalization and technologies such as RPA on organizational identity. Their study shows a shift towards more complex, knowledge-intensive work and the professionalization of the workforce. In particular, the study highlights how digital technology became a driver of this identity shift (see also Gomes and Seruca, 2023; Colombo and Beuren, 2023). Also, Kowalik et al. (2025) and Norzelan et al. (2024) outline how automation allows SSO staff to move towards higher-value activities. These studies portray a paradox of success: technologies that automate the original SSO mandate simultaneously undermine its cost-arbitrage model, triggering a pivot to new value logics.

Strategically, Wave 2 puts the spotlight on the issue of ambidexterity. As digitalization shifts value-creation from standardized execution to knowledge-intensive orchestration, SSOs must learn to explore new digitalization-driven opportunities while continuing to exploit at scale under the new conditions. This, however, is not yet addressed in the extant literature on digitalization in SSOs. In Kandampully et al.’s (2021) taxonomy, this second wave is best understood as service renovation: manual processes are automated, roles and skills are re-bundled, and new services are introduced. Where digital transformation extends to identity, for example, when an SSO morphs into a professional advisory hub, elements of reincarnation come into view.

Many scholars argue that companies must be efficient in managing today's business while also innovating and adapting to a changing environment. Since March's (1991) seminar article, a significant body of literature has addressed this tension between exploitation and exploration. While exploitation involves operational excellence, efficiency, and continuous improvement, exploration implies experimentation, variation, discovery, flexibility, and innovation. Successful companies are considered ambidextrous, able to combine both modes of operation. Ambidexterity is not straightforward, though, as exploitation and exploration carry divergent demands: While exploitation implies tight processes, performance management, and control, exploration requires looser processes, more autonomy, and freedom to experiment (Raisch and Birkinshaw, 2008). Building on this, research on service business development has shown that firms do not simply “balance” exploitation and exploration but enact them through different sensing–seizing–reconfiguring patterns (Fischer et al., 2010).

Scholars have suggested several ways of resolving this tension. Traditionally, multinational enterprises have pursued geographical ambidexterity (e.g. Ruano-Arcos et al., 2024; Khan et al., 2022) by leveraging location-specific advantages, allocating exploitative tasks to SSCs in low-cost regions, and retaining exploratory tasks in the parent organization. Alternatively, structural ambidexterity has been proposed, which separates these modes into distinct organizational units. These units encompass different competencies, and each can be tailored to meet the specific requirements of these two modes (Benner and Tushman, 2003). Some authors outline parallel structures where staff members oscillate between the two, depending on the requirements of the tasks at hand. This can imply a primary structure to address routine work, overlayed by a secondary structure of networks or projects that manage non-routine tasks and innovation (Adler et al., 1999). A different approach involves contextual antecedents, where leaders provide a suitable context that enables staff to choose for themselves how to switch between the demands of exploitation and exploration. In such a context, both strict performance requirements and trust are necessary to enable ambidexterity within a single organizational unit (Gibson and Birkinshaw, 2004). Orchestrating both demands is an essential dynamic capability of a service organization (Fischer et al., 2010). Yet, other authors emphasize the role of leadership in acquiring the capabilities needed for exploration while leveraging existing capabilities for exploitation (Volberda et al., 2001).

Although research on ambidexterity has grown substantially, its application to firms from emerging economies remains underexamined, with most studies focusing on established enterprises from developed markets (Hsu et al., 2013; Luo and Rui, 2009; Ruano-Arcos et al., 2024). Additionally, while scholars acknowledge the evolving nature of ambidexterity, limited attention has been given to how international organizations navigate the shift toward new forms, particularly in balancing global exploration with local exploitation (Ahammad et al., 2019; Turner et al., 2013; O'Reilly and Tushman, 2013).

SSOs provide an instructive context for advancing our understanding of these transformations in organizational ambidexterity, leading to our research question: How do SSOs as service providers transition from traditional exploitation to innovation-driven exploration activities, and what ambidextrous strategies, organizational dynamics, and capabilities enable this transformation?

We employed a multiple-case design (Eisenhardt, 1989; Yin, 2003) across nine SSCs in seven SSOs located in the Czech Republic, Singapore, Malaysia, India, and the Philippines. India, Malaysia, and the Philippines are top locations for SSCs globally, the Czech Republic is a top location in Europe, and Singapore is also ranked in the top 15 locations for SSCs by Kearney (2023), especially due to its good digital capabilities and favorable business environment. Two SSCs were chosen because they allowed us to visit a second SSC in another country for each SSO, enabling deeper insight into these SSOs. Using purposeful case selection (Morse, 2010, pp. 229–244) and our business partner's assessments, we targeted progressive SSOs that demonstrably lead in their industry, including the application of digital technologies, thereby establishing common boundary conditions (digital maturity; transformation intent). While three of the centers focused on accounting and finance, most interviews in the remaining centers also involved employees in these areas. The major method of data collection has been semi-structured interviews. The study draws on 38 in-depth, semi-structured interviews with 41 interviewees (four interviews involved two interviewees each, while in two of those, one interviewee participated in both). At each center, the center head was interviewed, followed by interviews with a range of team leaders. In three companies, one interview each with a representative of the retained organization was conducted, and one global shared services leader was interviewed. In addition, two interviews were conducted with two partners and one senior manager of an advisory firm that advises on shared services. See Table 1 for a list of interviewees. In our initial interviews, we asked questions that were quite broad, relating to recent trends and developments in strategy, operating models, and operations. As several interviewees spoke about transformative changes regarding digitalization and ensuing changes to their service offering and organization, we added questions on these themes to our interview guide to delve deeper into them in subsequent interviews. As interviewees spoke about the transformations that occurred and how they relate to their traditional operations, we followed up with more specific questions, such as: How do you organize the different types of tasks? Are they done by different people/teams? If so, are these teams organized and managed differently, and are they located separately? If not, how can people switch between different tasks? During the SSO interviews, we collected relevant background documents from the interviewees, including PowerPoint presentations on the centers' organization, their integration into the wider company, and the services they provide. Additionally, interviewees led us on guided tours of the offices, which we documented in visit protocols.

Working abductively, we iterated between emergent evidence and candidate lenses (Timmermans and Tavory, 2012). In our initial interviews, managers frequently mentioned how they explore new, higher-value, or transformation-related services while maintaining the current business's efficiency. These “structured contradictions” led us to consider the ambidexterity literature as a sensitizing lens, treating exploitation–exploration as a provisional theoretical vocabulary for what informants were describing.

Once we had settled on an ambidexterity framing, our analysis followed a theory elaboration approach (Fisher and Aguinis, 2017; Pratt, 2009; Dencker et al., 2023) to refine and extend the ambidexterity theory (e.g. March, 1991; O'Reilly and Tushman, 2013; Gibson and Birkinshaw, 2004; Raisch and Birkinshaw, 2008) in the context of service transformation (e.g. Kandampully et al., 2021). Theory elaboration is particularly suitable when a well-established theoretical frame exists but requires refinement to account for novel empirical settings, boundary conditions, or underlying mechanisms. The analysis followed an abductive process. Abduction is a mode of reasoning for explorative data analysis predicated on the interaction among theoretical frameworks as the interpretative lenses for analysis, case study data, and emerging theory (Maxwell, 2012, 2013; Orton, 1997). We used an iterative process of collecting, coding, and categorizing “empirical material as a resource for developing theoretical ideas” (Alvesson and Kärreman, 2011, p. 12) in constructing new theory. This entailed a process of subsequent abstraction from raw data through coding, categorizing, and linking categories to emerging themes and reflecting them with existing frameworks (Strauss and Corbin, 1998; Maxwell, 2012, 2013).

We began coding our data by following an open-ended, inductive discovery of emerging concepts within the data, which makes it possible to move from mass descriptive codes to fewer, conceptually abstracted codes (Strauss and Corbin, 1998). We adhered to the guidelines specified by Strauss and Corbin (1998) for constant comparison techniques and naturalistic inquiry, where the collection of data is iteratively intertwined with its actual analysis. The coding process of the transcribed interviews was supported by the software MAXQDA. After multiple rounds of cross-checking our draft with additional information from other data sources, we organized our results in the following coding structure (Gioia et al., 2013) (see Figure 1), the foundation of our case findings. We have structured the findings thematically, which allows us to integrate evidence from different sources, providing a coherent presentation (Miles and Huberman, 1994).

Initially, SSOs in our study were characterized by an exploitation of location economies and the performance of standardized, repetitive tasks. Then, service transformation introduced more complex, knowledge-intensive tasks. Two ways in which the simultaneity of old and new tasks was organized were identified: the “Moving up the value chain” and the “Transformation project” modes.

Since the 1990s, the seven companies in our study have each, at various points in time, established SSOs, each comprising several SSCs across the world in Europe, Latin America, and Asia to support functions ranging from HR to IT, Accounting and Finance to Procurement, and more. Initially, there was a clear split between activities in the centers, mainly focused on standardized, repetitive transactional tasks, and more complex tasks that were performed outside the centers, in the “retained organization” (RO), as a manager of a finance RO explained:

The SSO was regarded as that part of the company doing the simple tasks. Its managerial logic reflected the search for efficiency, predictability, and control, with sometimes low regard for people working there and the work done by them:

A key aim was to codify and standardize tasks to ensure that routinized service work yields predictable results. Using ERP systems, processes in the SSO were standardized to ensure that services would be of a consistent quality. Service workers followed predefined templates for action. As a manager explained:

For many tasks, deep professional knowledge was no longer required, as work was directed by ICT-driven standardized bureaucratic processes:

Overall, the initial situation in SSOs led to several consequences. While offshoring improved efficiency, predictability, and control of back-office service work, the high degree of standardization and control by machines over human behavior led to the deskilling of service work.

Due to the more recent digitalization, the situation has changed. We found that digitalization as a “path to service transformation” (Kandampully et al., 2021, p. 219) is eventually leading to the professionalization of service work in the SSOs investigated. At first sight, this seems paradoxical because digitalization makes even professional occupations like accounting, legal, and tax services increasingly vulnerable to being substituted by computerized routines (see Frey and Osborne, 2017). Increasingly, robotic process automation (RPA) has been automating mainly the less complex, more repetitive tasks in the SSOs in the study:

This posed a threat to the SSOs, whose existence depended on running these processes. So, a manager was stating:

And another manager asked:

The traditional, transactional tasks of SSOs were most vulnerable to automation. However, as our study indicates, rather than resigning and dissolving due to automation, SSOs reinvented themselves and ventured into more complex, knowledge-intensive, innovative, or transformational professional services. As several respondents pointed out, SSOs moved into expertise-based tasks, providing more value to the company:

To summarize, in the SSOs studied, digitalization in the form of RPA led to the result that rather than making SSOs redundant by automating their transactional tasks, SSOs in our study responded by moving to more complex, knowledge-intensive, or transformational tasks.

We found that in many SSOs in our study, the SSO overall was transforming by moving to a new level of more complex tasks, which we label the “moving up the value chain” mode. In others, the SSO was transforming by adding transformational tasks alongside the traditional tasks – what we call the “transformation project” mode. While in the latter case, the transformational tasks were connected to the traditional tasks, in the former, they were less connected. This had several significant implications for skill requirements, the organization of ambidexterity, and talent management. Figure 2 visualizes the antecedents and implications of these two approaches to the service transformation of SSOs.

In several companies, the SSO as a whole was moving to more complex, knowledge-intensive tasks, thereby moving up the value chain.

Connected to this, SSOs in the study have expanded their scope, adding more complex services that were not traditionally considered tasks for SSOs. This can include tasks, such as business partnering, that were traditionally seen as the remit of the retained organization, as they require developing a deep understanding of the business of the company and managing relationships with the business:

In one case, the SSO even took on a governance role, here for controlling, therefore for this task reversing the traditional division of labor, where now the SSO is setting governance standards that the remainder of the company has to follow:

In several cases, moving up means getting into direct contact with external customers and suppliers of the company, which was not the case in the past. Previously, the SSO staff was hidden away in the back office. This introduces a level of variation and requires flexibility from staff, two factors typically associated with exploration:

Conducting the more complex and knowledge-intensive tasks requires a new skill set and more highly qualified professionals. As the SSOs in our study were moving up to more complex, knowledge-intensive tasks, the nature of required professional staff was changing from lower-skilled employees to more highly skilled and experienced professionals, who have more career options, as outlined by various respondents:

Besides higher technical skills, professionals in SSOs are increasingly required additional, non-technical skills to master the new tasks, as one manager explained:

SSOs pursued two avenues to have more highly skilled staff. One of them was upskilling their existing staff:

At the same time, SSOs were also adapting their recruiting, now looking to hire new staff who already bring the higher skillset required for “moving up”:

At the same time, the more highly skilled staff expected a different work environment and talent management approach, as outlined above, that was more in line with “new work” rather than the more traditional factory-style environment typical of SSOs. This requires a “new logic” of talent management, as one SSC location head explained:

In terms of the organization of ambidexterity, companies that follow the “moving up” mode generally adopt a spatial separation approach. The SSO was increasingly shifting from its previous focus on exploitation to a new emphasis on exploration. Primarily during the transition phase, both modes coexisted side by side. As the whole SSO cannot move to more complex tasks from one day to the next, in these cases, a situation of ambidexterity occurred primarily during the transition phase, as some teams and individuals were already performing the new tasks, while others were still engaged in the more traditional tasks. Permanent new positions and teams were defined, or existing ones were redefined to enable the new, complex tasks. The new team performing the controlling governance mentioned above is a case in point:

Also, other SSOs set up separate teams for the more complex tasks:

One manager described how staff transitioned into the new teams. Staff from teams whose tasks were automated were moved into the new teams with the more complex tasks:

So, over time, new teams conducting exploration-type work were growing while old teams doing exploitation-type work were shrinking as the staff was shifted to the new teams and new staff was hired for the new teams.

In some of the SSOs in our study, the more complex and knowledge-intensive tasks had the character of transformation projects. They consisted of things like introducing new IT systems in the company, driving digitalization and automation, transforming processes, or implementing new management approaches such as agile organization. The SSO took on a pioneering and driving role for these innovations in the company, frequently piloting them in the SSCs and then rolling them out across the firm, as a team lead outlined:

In company G, for example, the SSO piloted agile ways of working and then led the rollout of them across other parts of the company:

Likewise, the SSO led the integration of the various – online and offline–channels by which customers connect to the company:

The SSO can play that driving role for such transformation initiatives, among other things, because it is one of the most global parts of the company, in contact with almost all units around the world:

A key difference between transformational tasks and moving-up tasks is that transformational tasks tend to be more closely connected to the traditional tasks the SSO has already been doing and continues to do. For example, when a new version of the SAP ERP system is implemented, SSO staff involved in the transformation benefit from having already been a power user of SAP all along in their routine tasks:

This affects the skill requirements for this category of more complex tasks: Rather than requiring new staff with a higher skill set or upskilling existing staff with professional training, the more experienced members of staff can be assigned to the transformational projects without extensive training requirements:

Team leaders then see these projects as stretch assignments that contribute to the development of employees:

The fact that staff going on transformation projects can then apply the new skills in their routine work, as in the second citation above, further demonstrates the connection between routine and non-routine tasks in this mode of ambidexterity.

Unlike tasks in the moving-up category, transformational tasks are added to the routine tasks and complement them rather than replace them. So rather than shifting towards new, more complex tasks, in this case, SSO staff oscillate between routine tasks and the added transformation projects:

From an ambidexterity perspective, the exploration (transformation) and exploitation (routine) tasks are organized in parallel structures, with staff switching between the primary organization and projects as a secondary organization. While the traditional routine tasks happen on an ongoing basis, transformation activities are organized as projects or in project mode:

Interestingly, the consequences of the transformation project mode for talent management are quite different from those of the moving up mode: While moving up led to new work-type talent management, allowing employees more freedom, transformation projects require more leadership touchpoints, that is, supervision, control, and support for staff. This follows from the fact that it is largely staff doing routine work that temporarily engages in transformation projects without having undergone significant upskilling. Therefore, the more complex and demanding transformation projects require closer involvement of leadership:

Our study into Shared Service Organizations (SSOs) undergoing service transformation through the lens of organizational ambidexterity has yielded several critical insights. Building on Kandampully et al.’s (2021) “3R” model of service transformation–reformation, renovation, and reincarnation–we find that the transformations observed across our cases align most closely with the renovation pathway. In line with this pathway, technologically driven change, most notably through robotic process automation (RPA), entailed the systematic dismantling of established services alongside the introduction of new ones, which enabled SSOs to move both rapidly and deliberately toward more complex service offerings and, in some instances, toward a more transformative role within their organizations. As Kandampully et al. (2021) note, such renovation is “executed by gaining sufficient knowledge, meticulous preparation, and strategically planning to enhance value for the customer” (Kandampully et al., 2021, p. 224). They further argue that renovation is likely applicable across a broad range of organizations and represents “the path with the greatest potential to transform firms' service to a higher level on a regular basis”. Our study advances this work by moving beyond the model's abstract articulation and showing, from an ambidexterity perspective, how service renovation of SSOs can be enacted through two distinct and empirically observable modes.

Drawing on an ambidexterity perspective, we observed that SSOs, traditionally typical exploitation cases, increasingly embrace exploration. This shift is intriguing, as it challenges the conventional roles assigned to SSOs, which are set up to run standardized, routine, repetitive processes most efficiently. This raises questions about how to organize this ambidextrous situation during the transition phase, when the old (exploitation) and new (exploration) logics coexist, with some parts already operating in exploration mode and others still in exploitation mode. Some authors have argued that – since these two logics require different types of organization and leadership – mechanistic vs. organic, hierarchy vs. heterarchy – it can be difficult to reconcile them within a single organization, while others outline ways of combining them (Raisch and Birkinshaw, 2008; Reihlen and Mone, 2012). In this context, we identified two distinct modes of ambidexterity: the transformation project mode and the moving up the value chain mode.

The key differentiator between these modes is the relationship between exploration and exploitation activities. There is an apparent oscillation between transformation projects (exploration) and routine activities (exploitation) in the transformational project mode. The organizing mode aligns with the concept of parallel structures, wherein a primary structure is optimized for executing routine activities efficiently, while a secondary structure—such as project-based initiatives—facilitates non-routine tasks and innovation. This dual-structure approach enables firms to simultaneously address the competing demands of exploration and exploitation within the same business unit (Gibson and Birkinshaw, 2004). This dynamic interplay became evident in the SSOs following a transformational project mode.

Notably, experienced and proficient staff members are involved in project-based exploration, dedicating their time alternately to activities of their routine task duties. These skilled professionals fluidly transition between periods of innovation—where they contribute to novel, complex projects—and phases of exploitation, where they assume roles to refine existing processes and enhance operational efficiency. More precisely than the notion of “parallel structures,” this dynamic aligns with what Adler et al. (1999) describe as “switching” — staff rotating between routine and non-routine tasks. However, this mechanism differs from contextual ambidexterity (Gibson and Birkinshaw, 2004), as staff members are assigned to non-routine tasks by their managers rather than deciding for themselves how to divide their time between exploration and exploitation. This may well be the case in the SSOs as staff, while upskilling towards more professional roles, can still not be compared to professionals in traditional professional service firms such as audit firms, who are in a better position to engage in self-directed switching, having, therefore, more autonomy over the allocation of their time between tasks (Nordenflycht, 2010; Mintzberg, 1979).

The moving up the value chain mode reflects a lasting shift from routine to complex activities rather than switching back and forth. In this mode, SSOs gradually shift from exploitation, focusing on efficiency and performance in their existing operations, to greater exploration to seek knowledge-intensive growth opportunities. In this mode, ambidexterity was typically organized through a structural or spatial separation of units within the SSO, with some focusing on exploration and others on exploitation (O'Reilly and Tushman, 2004). This is also consistent with what Adler et al. (1999) call “partitioning” – having separate subunits for routine and non-routine tasks. What is specific and interesting in our case, though, is that over time, more and more staff who can be upskilled move into exploration-oriented teams as their routine tasks get automated, joining newly hired staff who already have the required skill set. Thereby, in the SSOs in our study, the spatial separation is not static. The boundary between the two types of teams is shifting as exploration-oriented teams grow while exploitation-oriented teams shrink.

Across both modes of ambidexterity, it can be noted that the boundary between exploration and exploitation that used to run between the SSO (doing exploitation) and the retained organization – those parts of a function that have not been moved into the SSO – (doing exploitation) – has now moved into the SSO, thereby creating the need for ambidexterity within the SSO. Moreover, we outline which antecedents will encourage SSO executives to pursue the “Moving up the value chain” or “Transformation project” mode of ambidexterity to address the conflicting demands of the old and new service offerings. When service renovation took the form of adding transformation tasks related to the existing tasks, they could be performed by existing experienced staff, with supervision as a development opportunity. This encouraged the transformation project mode of ambidexterity. To the contrary, when service renovation involved introducing more complex tasks that were less related to existing tasks, they required more highly skilled staff and encouraged a moving-up-the-value-chain mode of ambidexterity, setting up permanent new positions and teams distinct from existing ones. These can be managed in ways that are better suited to more complex, knowledge-intensive work and the expectations of the required, more highly skilled staff (see Figure 2).

Our findings also contribute to research on leadership in SSOs as they become more like professional service firms (PSFs). It has long been known that the management of professionals faces the challenges of their high need for autonomy, which has been likened to “cat herding” (Nordenflycht, 2010). Professionals are highly qualified and often have multiple options in the labor market, requiring special attention to engage and retain them. This requires a leadership style that includes providing autonomy, delegation, flexibility, career development opportunities, and opportunities for further professional education (Empson and Langley, 2015; Empson, 2017; Smets et al., 2012). In the SSO context, however, professional work has originally – as outlined above – been highly rationalized and reduced to standardized processes, which often led to the recruitment of non-professional staff or only junior staff that would sooner or later leave the SSO. Therefore, work, staff, and talent management in SSOs have traditionally differed from traditional professional work, often resulting in “demeaning experiences” (Lepistö et al., 2018) for service workers. Now, with the newly emerging roles of SSOs, new talent requirements, and new talent management, professional work and management are becoming more similar to traditional professional work settings and management outside of an SSO environment. The implication is that what we know about the management of professionals in conventional environments is also becoming increasingly relevant to professional work in SSOs that are repositioning themselves by moving up the value chain or engaging more with transformational projects.

Yet, the challenges for leaders in the two types of ambidexterity are quite different. Managing the parallel structures of transformation projects vs. routine work (switching) requires planning, organizing, and leading two very different processes (exploration and exploitation) simultaneously within the same organizational structure, creating high coordination demands. In particular, balancing the inherently conflicting goals of assigning more senior, experienced staff to temporary project work while ensuring smooth routine operations creates tensions that leaders need to anticipate. Managing the structural ambidexterity that comes with the “moving up” mode (partitioning), on the other hand, requires managing two largely separate operations that run in parallel but gradually shift to take in more knowledge-intensive work. Managing tensions across these units as they operate with very different organizational logics – the exploration unit following a professional logic (Nordenflycht, 2010), while the exploitation unit reflects the McDonaldization of service work (Ritzer, 2011) – becomes a significant managerial challenge. In summary, leading ambidexterity in SSOs revolves around managing transitions, balancing conflicting demands, and fostering adaptability over time. Leaders must embrace the paradox of being efficient and innovative and simultaneously entertain an ambidextrous organization.

In the context of shared services, the ambidextrous lens is a useful framework, as it helps recognize specific constellations of exploration and exploitation activities as these organizations undergo service transformation due to the disruption of digital technologies. Thus, understanding and managing ambidexterity within SSOs as they transition towards a new organizational form remains a compelling subject due to its inherent paradox and complex challenge. We hope that this research stimulates further debate in this endeavor.

Managers of Shared Service Organizations need to think carefully about how to organize the transformation of their SSOs as they respond to disruptive technologies such as RPA and AI. Our findings offer practice-oriented implications. First, match the ambidexterity design to task-relatedness: when new services are adjacent to the legacy portfolio, so they can be performed by existing experienced staff, our study suggests using a transformation-project mode. Yet, when new services are rather distant and require a higher skill set, adopt a moving-up-the-value-chain mode. Second, our study indicates the need to professionalize the talent model. That requires managers to upskill incumbents, hire new talent, redesign roles and career ladders toward more knowledge-intensive work (e.g. analytics/advisory tasks), and align rewards.

According to our study, the transformation project mode can pose challenges when projects need to be extended, and staff seconded to a project full-time are absent from their routine tasks for longer than planned. This requires careful contingency planning as unplanned delays are not uncommon on projects. In SSOs where the various SSCs in different countries are well integrated, staff from SSCs in other countries might be able to cover. In the moving up the value chain mode, we encountered the challenge of SSOs no longer being recognized in the company as the “cheap back-office operation”. In SSCs in Asia, this legacy may be compounded with cultural limitations on self-confident appearance. To obtain recognition, providing training for SSO staff on an assured presence, professional certification for staff, and a new branding for the SSO can help.

Our findings should be interpreted considering the following limitations. First, the sample is regionally concentrated: nine SSCs located in Eastern Europe and Asia. SSCs in Latin America, Africa, and other jurisdictions may operate under different business systems (labor markets, regulation, digital infrastructure) (e.g. Whitley, 2007). Comparative work that varies across country–industry–institutional conditions could elaborate on and test the transferability of the mechanisms we identified and surface alternative pathways of renovation or reincarnation (Kandampully et al., 2021). Future research might also explore whether SSOs embedded in more fluid, ecosystem-like service arrangements of the future show different participation, governance, and value-capture logics than the SSOs we studied (Subramony et al., 2018). Second, we purposefully sampled advanced SSOs to explore transformation-related phenomena, thereby foregrounding “frontier” practices but underrepresenting lagging organizations. Studying late adopters, including the challenges they face in governance, skills, and legacy architectures, could reveal distinct adaptation logics (e.g. imitation, bricolage, defensive standardization) and different ambidexterity designs. Third, while our multiple-case design helped us identify essential patterns across cases, case triangulation and comparison were limited by data constraints. Fine-grained, in-depth, ethnographic case studies (e.g. Ybema et al., 2009) would illuminate how managers and employees appropriate RPA/AI, and how affordances emerge and stabilize (Leonardi and Vaast, 2017).

We would like to thank editor Linda Alkire and the two anonymous reviewers for their constructive guidance. We also thank the anonymous reviewers of the AOM Annual Meeting 2024 for their helpful comments on an earlier version of this paper.