Integrating advanced digital technologies in existing lean-based production systems: analysis of paradoxes, imbalances and management strategies

The advent of advanced digital technologies, typified by Industry 4.0, opens up new and innovative opportunities for manufacturing companies to enhance their production systems through connectivity and intelligence (Calabrese et al., 2020; Xun et al., 2021). In this context, many companies are in the process of transforming their operations to smart production (SP) to achieve sustainable and competitive manufacturing practices (Li, 2020; Sebastian et al., 2017; Sousa-Zomer et al., 2020). To enable SP, the required advanced digital technologies need to be carefully planned and integrated into existing production systems (Fattouh et al., 2023; Frank et al., 2019; Leberruyer et al., 2023). During the last few decades, many production systems have been based on Lean production (LP) principles and now face a new paradigm shift to achieve SP (Alavian et al., 2020). This integration process is often characterised by ad-hoc and fragmentary initiatives (Zangiacomi et al., 2020; Björkdahl, 2020), driven by different organisational functions within the company with varying purposes and motivations or distributed production teams located in various worldwide geographic locations (Ahlskog et al., 2023).

Hence, the difficulty of integration is often not the digital technologies themselves but the challenges in changing from a foundation in the “physical world”, to the tensions that arise from interweaving the physical with the digital world (Kane et al., 2015; Saabye et al., 2022). Therefore, to leverage the potential of digital technologies into existing Lean-based production systems (LPSs), manufacturing companies need to redefine and reconfigure organisational strategy, structures, infrastructure, resources and culture (Chirumalla, 2021; Putra et al., 2023). In this regard, recent studies in operations management (OM) have proposed utilising a socio-technical perspective incorporating workers and organisational issues to shape the integration of digital technologies aligned to LP (Marcon et al., 2022; Rosin et al., 2020; Tortorella et al., 2019). According to Buer et al., (2021), the simultaneous use of Lean and digital technologies has a greater complementary effect than the individual effects of each approach. The study suggests that rather than being obsolete, LP is still crucial in effectively utilising emerging digital technologies. Several researchers have also emphasised that LP serves as the foundation for implementing digital technologies (e.g. Rossini et al., 2019; Sony and Naik, 2020). The absence of proper coordination and alignment between the two can lead to multiple tensions during the implementation process (Pozzi et al., 2023; Mokudai et al., 2021), e.g. the tensions between centralised digitalisation of processes vs. decentralised improvement work (Holmemo and Korsen, 2023. Tension in this context refers to conflicts or unease that arise when trying to harmonise disparate elements (Putnam et al., 2016). For example, a tension may arise when LP prioritises minimising waste and keeping processes simple, whilst digital technologies introduce complexity and data-intensive processes.

Paradox theory is well suited to addressing tensions between seemingly contradictory approaches (Smith and Lewis, 2011; Putnam et al., 2016) such as LP and digital technologies because it encourages a balanced, dynamic and learning-oriented perspective (Smith and Lewis, 2011). Rather than attempting to eliminate the tensions, it encourages organisations to harness them as sources of innovation and competitive advantage (Raisch et al., 2009). However, the existing literature provides limited insights into the emerging paradoxes associated with integrating digital technologies into LPS (Holmemo and Korsen, 2023; Putra et al., 2023). Researchers have explored organisational paradoxes, underlying tensions and potential management strategies within the context of LP implementation (Maalouf and Gammelgaard, 2016; Erthal and Marques, 2022) or advanced digital (or industry 4.0) technologies implementation (Dieste et al., 2022; Smith and Beretta, 2021; Mishra et al., 2022; Tóth et al., 2022). Nonetheless, a notable gap exists in respect of research that examines the paradoxical tensions arising from the combination of LP and digital technologies, creating a gap in theory that this study aims to fill. Although studies have emphasised the importance of the interplay between LP and digital technologies (e.g. Rossini et al., 2019; Sony and Naik, 2020; Pozzi et al., 2023), there is a lack of micro-level understanding of paradox resolution strategies enabling their integration, considering factors such as temporality (e.g. Poole and Van de Ven, 1989) or organisational ambidexterity (e.g. Secchi and Camuffo, 2019) and how these factors underlie paradoxical tensions.

Therefore, the purpose of this paper is to advance the understanding of paradoxes, underlying tensions and potential management strategies when integrating digital technologies into existing LPS in a manufacturing industry context, with aim of achieving synergies and fostering the development of LPS. More specifically, two research questions are posed:

This study adopts a collaborative management research (CMR) approach involving two global manufacturing companies to gain insights into the paradoxical tensions arising when integrating digital technologies into LPSs. In identifying aspects of performing, organising, learning and belonging paradoxes (Smith and Lewis, 2011), we found that resolving one tension might unintentionally create new ones (Poole and Van de Ven, 1989). For instance, separating LPS- and SP-teams manages the performing paradox but this triggers an internal drift rendering a belonging paradox. Consistent with the dissipative equilibrium model (Weiser and Laamanen, 2022), our findings showcase the tendency of internal drift and the forces that constantly push the organisations towards imbalances, which in worst case tilts the organisation and generates vicious cycles constraining any integration of LP and digital technologies (Smith and Lewis, 2011; Holmemo and Korsen, 2023). Thus, the study highlights the sensitivity and temporality of equilibrium states for all paradoxes (Poole and Van de Ven, 1989), and the findings suggest that maintaining dynamic equilibrium requires new strategies like synthesis (Dieste et al., 2022) and micro-shifts of re-contextualising to rebalance paradoxes. Paradox resolutions are as such context-sensitive and system dependent (Schad and Bansal, 2018). Our study confirms the complementary effects of LP and digital technologies (Buer et al., 2021), however, we also identify contradictions that appear in the practical implementation leading to paradoxical tensions (Maalouf and Gammelgaard, 2016; Dieste et al., 2022; Frank et al., 2019; Smith and Beretta, 2021; Erthal et al., 2021).

The integration of digital technologies with LP has emerged as a significant area of interest within OM. When reviewing research on tensions in integrating digital technologies and LP, three clusters of papers emerged as shown in Table 1. The first cluster of papers investigate tensions that arise when implementing LP by utilising a paradox perspective whilst not considering perspectives of any digital technologies (Maalouf and Gammelgaard, 2016; Erthal and Marques, 2022; Secchi and Camuffo, 2019). This small set of papers focussed on case-based research and found that organisational culture and the way of organising Lean practices in the process are critical. The second cluster of papers embrace papers on tensions in implementing I4.0 by utilising a paradox perspective, whilst not considering perspectives of LP (Dieste et al., 2022; Putra et al., 2023; Smith and Beretta, 2021; Mishra et al., 2022). These studies are majorly based on single cases and proposed resolution or coping strategies to manage diverse set of paradoxes. Whilst not using paradox theory explicitly, Holmemo and Korsen’s (2023) paper distinguish itself from the other papers by also identifying paradoxical relationships between lean and digitalisation.

The third, and largest cluster of papers, is focussing on the integration of I4.0 and LP, but do not explicitly utilise a paradox perspective (Rosin et al., 2020; Sanders et al., 2017; Tortorella et al., 2019, 2021; Buer et al., 2021; Rossini et al., 2019; Mokudai et al., 2021; Holmemo and Korsen, 2023; Kumar et al., 2023; Von Haartman et al., 2021; Bittencourt et al., 2021). Most of these studies are dominated by survey method, compared to case studies, which is commonly used when exploring paradoxes in the integration of either I4.0 or LP. These survey studies confirmed that there is an interesting synergistic relationship between I4.0 and LP concepts and in combination they have a better effect that can lead to improved operational performance. Moreover, LP is considered as a foundation for successful implementation of I4.0. One plausible explanation to a successful integration of I4.0 and LP is that companies need to have management engagement and improvement processes in place (Bittencourt et al., 2021), which implies the importance of an improvement culture and that improvement initiatives of various kinds, in which digitalisation is one source for improvements, needs to be balanced. Hence, a mature LP approach seems to create a readiness for I4.0 integration (von Haartman et al., 2021). However, there are less tensions or contradictions identified in the survey studies of I4.0 and LP. In the two case studies, contradictions are more salient. For instance, Mokudai et al. (2021) describes what they refer to as a lean trap, i.e. when problems and potential digital technology solutions are defined by people operating within the current LPSs, this might limit the view of, and interest in, more radical and disruptive solutions. In a similar vein, Holmemo and Korsen (2023) identify the lean-digitalisation paradox providing an explanation for the challenges in integrating I4.0 and LP by asserting that digitalisation is agile, yet centralised and programmatic and that digital tools are standardised yet locally optimised and asynchronous.

When new elements, such as digital technologies, are intended to be integrated into existing LPSs, differences in intents and priorities generate tensions that can be perceived as contradictory, thus leading to what have been framed as organisational paradoxes, which in one way or the other need to be managed to positively influence organisational development (Putnam et al., 2016; Dieste et al., 2022). Organisational paradoxes are constituted of contradictory yet interconnected elements that exist simultaneously and persist over time (Eisenhardt, 2000; Smith and Lewis, 2011). One example of a well-known paradox in organisations is the one between the two modes of organisational learning exploitation (i.e. activities conducive for short-term efficiency) and exploration (activities conducive for a long-term effectiveness and innovation) (March, 1991).

Smith and Lewis (2011) distinguish between four paradoxes commonly observed in organisations. Performing paradoxes arise from the need to meet diverse stakeholder expectations, resulting in competing strategies and goals, concerning the plurality of organisational goals, where stakeholders may have divergent views on what constitutes organisational success, such as financial goals versus social goals. Organising paradoxes stem from organisational systems, leading to competing designs, structures and processes, hence, they revolve around different approaches to structuring work, such as centralisation vs. decentralisation or standardisation vs. customisation. Belonging paradoxes occur when individuals seek both self-expression and group affiliation, creating tensions between personal identity and collective relationships and involves navigating between the individual and the collective. Learning paradoxes highlights the balancing act between existing knowledge (exploitation) and new ideas (exploration) i.e. the need to build upon existing knowledge and practices whilst also embracing innovation (Ozanne et al., 2016).

Apart from the theoretical recognition of organisational paradoxes, these paradoxes also need to be accepted and embraced by actors in an organisation because failing to manage paradoxes can lead to either stagnation or fragmentation (Dooley and Van de Ven, 1999; Raisch et al., 2009). The first step of managing or resolving paradoxes is making sense of and accepting a paradoxical tension, which means to recognise that the two sides/elements are simultaneously conflicting and interconnected and therefore that they require a comprehensive and joint solution. Acceptance also implies an awareness that prioritising one element over the other will create new issues elsewhere or later with possibly increased magnitude and range (Poole and Van de Ven, 1989; Smith and Lewis, 2011). Once a paradox is accepted and embraced, different type of actions and strategies to manage or resolve the paradox need to occur to create or maintain the dynamic equilibrium of the paradoxical tension (Smith and Lewis, 2011; Dieste et al., 2022; Weiser and Laamanen, 2022), which also has been the main focus in theory on organisational ambidexterity, i.e. the organisation’s ability to manage the different poles of a paradox simultaneously (Raisch et al., 2009; Secchi and Camuffo, 2019).

Previous research has identified different types of resolutions strategies to manage the juxta-positioned poles of a paradox simultaneously (Dieste et al., 2022). One type is structural separation (cf. structural ambidexterity) and refers to either 1) spatial separation, which means placing the paradoxical elements at distinct places or levels of context, or 2) temporal separation, i.e. allocating them in diverse time periods. Whilst other strategies are aiming for contextualisation as a strategy. Contextualisation is mainly focussed on the cognitive or behavioural aspects (cf. Raisch et al., 2009 on contextual ambidexterity) and helps managers to develop the capability to think paradoxically (Smith and Lewis, 2011; Xiao et al., 2019). because it (1) tries to soften the paradoxical tensions and make them more “workable” and (2) helps in “alleviating tensions by identifying relevant stakeholders as well as their aims and to define how they can be integrated to address the tension” (Dieste et al., 2022, p. 3). To enable contextualisation, the managerial practice of facilitation of learning has proven to be a key factor (Maalouf and Gammelgaard, 2016). A third strategy is synthesis and involves introducing new elements (such as novel organisational procedures and routines or technological solutions) that can reconcile both poles of the paradox (Dieste et al., 2022).

The three types of resolution strategies correspond to what Weiser and Laamanen (2022) frame as “pre-emptive balancing” in their conceptual dissipative equilibrium model, which are the strategies used by management for splitting (i.e. separation strategy) and integrating (contextualising strategy) the poles of the paradox. However, another critical dimension of managing organisational paradoxes advocated by Weiser and Laamanen’s (2022, p. 8) model is that “the organisational system is continually pulled out of equilibrium by organisational and human tendencies to drift, as well as by the unintended consequences of interactions between actors”. As such, the dissipative equilibrium model also views paradox management also as a continuous process of navigating paradoxical tensions through micro-shifts, requiring apart from the pre-emptive balancing of managers, continuous corrective and proactive actions to maintain organisational balance.

This study is based on CMR (Shani et al., 2008) involving two global manufacturing companies. The study seeks to highlight the paradoxes, underlying tensions and potential management strategies when integrating digital technologies into existing LPSs, with the aim of achieving synergies and fostering the development of LPS. CMR is well suited for the involvement of multiple perspectives and expertise (Coughlan et al., 2016), which by providing field presence (Langley and Klag, 2019) and an insider’s gaze (Brannick and Coghlan, 2007), is crucial for understanding and addressing the intricacies of this integration. CMR is particularly valuable for both enhancing practical applications and providing valuable contributions to theoretical frameworks in the field of OM and organisation studies (Shani et al., 2012). Further, it enables a holistic and comprehensive understanding of tensions and organisational paradoxes (Shani et al., 2008). Thus, the possibility of obtaining versatile and complementary insights is opened, which is critical since the integration of digital technologies into LPSs is a complex and multifaceted challenge that involves various stakeholders at both the strategical and operational levels.

We applied two cases to collect the relevant data. Based on current practice (Goffin et al., 2019), we opted for theoretical sampling to select cases that would provide insights into the integration of digital technologies into existing LPS. The selection of the cases was informed by the overall CMR approach and the study´s purpose, which highlights the following aspects: the relationship between the researchers and the practitioners, the existing LPS and the strategies and organisation for integrating digital technologies.

First, we selected two global manufacturing companies (Table 2), Scania and Volvo Construction Equipment (Volvo CE), each leading within their respective market segments, with which well-established and long-term collaboration between researchers at the university and two managers (PhDs) already existed. The two managers, expected to contribute to the collection of a rich dataset and help in providing valuable insights into industry and indeed they actively contributed to the framing of the research focus and the study design (Shani et al., 2012). All five authors actively participated in the various phases of the study, including data collection, thorough data analysis, collaborative interpretation of the findings and the subsequent identification of potential action steps. The two managers at the companies were engaged as insider-researchers (Brannick and Coghlan, 2007) and as they were “native” to the organisations they were instrumental in helping to identify diverse stakeholders who could provide rich descriptions of the integration of digital technologies into the existing LPS and provide in-depth information about its key activities, supported by documents and background information. This provided the opportunity to contrast various perspectives on the integration of digital technologies into an existing LPS.

Second, we considered it as important to select cases where there was longstanding experience of working with LP. Both companies possess over 2 decades of experience in working with LPS and have been highly successful in their implementation of LP capabilities for improvement of the LPS. Thus, a key selection criterion was the maturity of the companies in their Lean journey.

Third, both companies continuously work with LPS-development and considered it as critical to adopt and integrate advanced digital technologies in finding new competitive and sustainable production approaches. Further, both companies have implemented actions dedicated to digital technologies at both company and factory level with well-defined strategies and organisations and both have high expectations about to be derived from a transformation to SP. Accordingly, both companies had already gained experience in integrating digital technologies into their existing LPS.

With a shared interest in improving comprehension of the encountered challenges, the managers played an active role in both planning and executing the study (Shani et al., 2012). Throughout the process, the team maintained ongoing interaction, including face-to-face and virtual meetings, as well as regular email exchanges. The study, which forms the basis of this paper started in autumn 2022. Data was primarily gathered in in-depth interviews with key respondents, supplemented by case documents, two focus group sessions and regular research team meetings where the expertise of the two managers was drawn upon. This approach was well-founded due to the extensive collaborative history and deep knowledge concerning LP and digital integration within the two companies. This is especially true when working with a relatively focussed and uniform population that aligns with the study’s purpose (Hennink and Kaiser, 2022). The 11 respondents (Table 2), selected in accordance with purposeful sampling principles (Patton, 2002), provided comprehensive insights into both the strategic and operational aspects of LP and digital technologies integration. The in-depth interviews were conducted solely by university researchers, without active manager involvement. To ensure a structured approach for the interviews, an interview guide was developed, focussed on three main themes:

1. The development process of new technologies and digital services, including questions such as: How are SP-team organised and who are your stakeholders? Which digital technologies do you usually work with SP-team? Describe the typical development process? What are key stages and decision points (Start/stop/continue)? Which functions are involved in the development process?

2. The relationship between digital technologies, SP and LP, including questions such as: Which, if any, digital technologies seem to be more prominent in enhancing LP? Why? What experience have you gained/what lessons have you learnt from deploying digital technologies in LP?

3. The integration and implementation of new technologies within the organisation, including questions such as: How is the handover of technology to a recipient/another organisation handled? How is existing experience and knowledge preserved/supplanted/amended in the receiving organisation due to the new solution?

Additionally, two focus groups sessions were organised, one at each company. In total, six respondents participated in the two sessions. The dual aim of the focus group interviews was to gain more detailed understanding and further explore tentative findings derived from the in-depth interviews. In the focus groups interviews the respondents took part of tentative findings and illustrations and in both sessions, there was a strong recognition, i.e. gain face validity, amongst the respondents an indication that we had reached saturation in terms of enough data to underpin our conclusions.

Data analysis was a collaborative effort involving all authors. Findings were placed in “collective inquiry” where the tensions of the integration of digital technologies were investigated. Inclusion of the two in-case managers, with their first-hand experiences and expertise within the companies, proved invaluable, providing a deep contextual understanding of the respondents’ statements and contributing to a more nuanced interpretation of the practical examples shared by the respondents.

Literature analysis commenced using the four paradoxes performing, organising, belonging and learning as preparatory concepts to provide direction in looking for patterns in the data. Consequently, our comprehension and utilisation of the concepts evolved concurrently with data collection and analysis through a structured and deliberate process of bringing together different concepts. These concepts were not explicitly introduced to the respondents but were employed to aid the researchers in theory development. Our data underwent thorough analysis based on the guidelines for thematic data analysis (Miles et al., 2018) encompassing three simultaneous activities: data reduction, data display and conclusion drawing. Initially, the interviews were recorded and transcribed, allowing us to become thoroughly familiar with the data. Each researcher read every transcript repeatedly, identifying interesting phrases that were transferred into a spreadsheet. This process facilitated the identification of common constructs and potential relationships through cross-interview analysis (Eisenhardt and Graebner, 2007).

Subsequently, we applied concepts from the analytical framework (Weiser and Laamanen, 2022) to analyse: characteristics of the integration of digital technologies; acceptance of organisational paradoxes; searching for tendencies to drift and related tensions; and searching for patterns of corrective and proactive actions/strategies. The analysis resulted in first-order categories of codes that represented respondents’ perspectives, expressed in their own words, phrases, or terms (Miles et al., 2018). These provisional codes served to organise subsequent data whilst remaining open to the identification of emerging codes. As a result, the emergence of themes was driven by both theory and data. Next, the analysis focussed on identifying links and patterns by comparing and combining the first-order categories, which lead to the formulation of second-order themes, including a list of initial codes divided into categories and created preliminary drafts illustrating how the categories related to the theoretical themes. This process involved a thorough exploration of the meaning, distinctions and commonalities of codes and categories amongst the authors, culminating in the development of a draft thematic map, representing the different types of tensions involved related to the four paradoxes proposed by Smith and Lewis (2011), Finally, we conducted a comprehensive review of the labelling of codes, categories and themes to ensure their clear definition, designation and alignment.

To enhance the validity of the research, the author team including the two managers had extensive discussions on the initial conceptual framework and subsequent empirical findings. The CMR approach allows for structured exploration and exploitation by practitioners and researchers with the intention of reducing the likelihood of drawing false conclusions from the data collected (Coughlan et al., 2016). To minimise observer bias all interviews were done by at least two of the researchers and the analysis was done jointly (Eisenhardt and Graebner, 2007). Although the main focus of the research has been to identify patterns concerning paradoxical tensions rather than validating them, the topic is rather unexplored and thus the pattern matching was difficult and can have impacted the validity negatively (Denzin and Lincoln, 1994). To mitigate the negative effects of conducting individual interviews, we completed them with focus groups to validate preliminary findings and gain face validity of the identified paradoxical tensions from the individual respondents. Further, to strengthen external validity data were collected from multiple respondents representing different internal functions as well as different sources of evidence aiming to triangulate responses and data. For example, we compared tensions identified from the interviews with governance models (e.g. technology roadmaps) of the respective company. At the start of each interview, we explained the paper’s idea, clarified key concepts to minimise the risk of misunderstandings.

This study took place in the context of manufacturing operations in two companies: Scania and Volvo CE. Based on the principles of LP and inspired by the Toyota Production System, both companies have over the years gradually developed their own company-specific LPS (see Table 2), which prescribe and standardise the best practices of daily production in manufacturing operations, as well as providing a guide on how to work on short-term (continuous improvements) and long-term production system development.

Over the last five years, a critical issue has been how to increase the utilisation of advanced digital technologies and how the current LPS can be transformed into a SP system. That is, the cases are not about building a Smart factory from scratch, rather it is about finding ways to gradually implement new digital technologies and make them fit with already established production settings to contribute to an overall efficient and effective LPS. Both companies utilising a future strategy when making investments in new machinery by ensuring that they are “IoT ready” for potential future needs, even if currently they have not yet identified the possibilities or even the needs. At Scania, examples of digital tools already applied or explored are visual aids for the assembly workers such as smart glasses that provide information in real time, digital team boards allowing for better storage and retention of production data enabling better analysis of trends and patterns in operations and enhanced systems for data driven maintenance. All these solutions are intended to increase productivity, enable problems-solving and control variation in the process in line with the intentions of the LPS. At Volvo CE, digital tools and technologies applied or explored are, for example, digital Lean boards to support real-time updates for daily production meetings and to obtain real-time status of the factory. Other technologies are brownfield connectivity to connect old machines/equipment, vision systems to automatise quality control, exoskeletons to support heavy lifts, additive manufacturing for tools and fixtures, XR-technologies to facilitate remote digital collaboration and collaborative robots to remove repetitive tasks for humans.

To support the work on the exploration and implementation of digital technologies in production, both companies have established a type of specialist support function, referred to as a “smart production team” (SP-team), with the aim being to increase the organisational performance in manufacturing operations over the medium to long term based on a set of strategies and technology roadmaps. The SP-teams were created to explore new information about advances in digital technology and develop competence in digitalisation and future manufacturing technologies. When first created, the SP-team at Scania used a top-down approach because they were pushing new digital technologies at production, whilst to begin with, Volvo CE adopted a bottom-up approach in which the objective of exploring and implementing digital technologies was to accelerate the company´s LP journey, whilst at the same time adapting to the future by taking advantage of the digital transformation. Later, however, new approaches evolved and now in both cases the mission primarily to inspire the production units and sites and support them in how to apply and utilise digital technologies.

In both companies, the integration of digital technologies into the existing LPS has been associated with many competing demands and prerequisites that need to be prioritised. Hence, the work on integrating digital technologies is characterised by a continuous process of negotiation and compromise between the goals, needs and requirements of multiple stakeholders. For example, when innovation engineers responsible for the integration of digital technologies are working on testing and implementing new sensors for real-time monitoring of an assembly line with the long-term goal of reducing disturbances and waste, this presupposes a disruption of the daily production and interference with the day-to-day needs of production managers and operators, who are under pressure to maintain a steady production of goods and who must be ready to manage disturbances to avoid unplanned disruptions. Compared to the work of the LPS-team and SP-teams, the work of managers and operators in production is in general much more short-sighted focussing as it does on operational performance and key performance indicator's (KPI’s) related to the day-to-day production. Hence, there is a tension between long-term goals and short-term goals. Furthermore, a key characteristic of both companies, is that the senior production managers have a strong level of autonomy, which means that they can largely avoid or ignore any “push” from the support functions or other instances and decide autonomously when to implement which new ways of working, including new technologies.

Even though it shares the goal of improving the operational performance of the LPS, the integration of digital technologies in the existing LPS is thus characterised by multiple contradictory goals and demands, different time perspectives, changes to requirements, etc. that arise at the intersections between different stakeholder interests, such as between production, SP-team, LPS-team and senior management.

Previous research provides evidence of the potential synergies between LP and digital technologies for production system development (Buer et al., 2021). In this section we address RQ1 and by adopting a paradox theory (Smith and Lewis, 2011) perspective, it allows us to a get a more detailed understanding of tensions and related paradoxes (see Figure 1) that first need to be accepted and embraced and then secondly be pre-emptively balanced and work through (Weiser and Laamanen, 2022) to enable the integration of digital technologies in LPS. Illustrating quotes of paradoxes are also presented in Table 3.

In the analysis, it becomes apparent that the goals and intentions of the different stakeholder interests in manufacturing operations exhibit differences, both in terms of scope and time horizons, generating the paradox of current goals vs. future goals. For example, the production units, with their decentralised authority, operate under one performing regime with a strong emphasis on optimising day-to-day production of goods, which are at variance with the long-term goals of the SP-team. Additionally, many respondents also point to the “senior management” level, highlighting the need for shared understanding of goals or time horizons between LPS-team, SP-team and production units. This potential disparity between senior management and the other stakeholders establishes a second performing regime.

A second performing paradox concerns competing but potentially synergetic approaches to LPS-development. In the interaction between goals related to LP and goals related to digital technologies two distinct performing regimes can be identified, that generate tensions due to contradicting priorities. SP-teams are focussed on exploring new technologies and capabilities, which contrast with the incremental improvement approach of already established production technologies favoured by the LPS. To manage the paradox structural separation into different support structures (LPS-teams and SP-teams) then becomes a solution that allows the companies to withhold focus on both issues and instead build expertise competence. However, the structural separation triggers an internal drift of the belonging paradox. To achieve synergies between the two perspectives, there is also a need for mutual understanding of one another’s goals, approaches and potential benefits, a fact that is illustrated by one of the SP managers at Volvo CE: “We need to recognize a problem first and then reflect about the methods that we have in Lean today can solve this problem, but also to look at digitalization and what it could help to gain even greater results”. Cross-functional work across specialist functions and production then becomes a key to the synthesis-resolution strategy. The four performing regimes embody different underlying motivations that are contradictory and generates tensions, yet need to be interconnected as part of the performing paradoxes to positively influence the LPS as a whole.

Both companies express a concern about incorporating digital technologies in ways that might introduce inflexibility and constrain future improvement work. To avoid built in inflexibility in the integration of new technical solutions the companies try to manage the paradox of pull based on needs vs. push based on opportunities. Both companies emphasise the significance of developing both need and opportunity-based solutions that are easily comprehensible to the users, such as operators and technicians. It is crucial for the end users of these solutions to have confidence in their effectiveness, allowing for their integration into the LPS. This can be achieved through either push or pull mechanisms. According to the respondents, a need-driven pull-approach allows them to address the correct problems with the most suitable short-term solution, whether it be a digital solution or some other form of solution. Adopting a push-approach typically requires a long-term perspective because an increased awareness of possibilities must trigger the will to invest time and resources.

Related to the paradox of pull vs. push, is the paradox of using LP principles vs. implementation of new technology. Respondents from Volvo CE have previously highlighted the importance of introducing new technologies and have performed the necessary work to understand the root cause of the situation and, as such, utilise LP principles before subsequently incorporating the relevant digital technologies. For example, it was discovered that in certain cases, the most effective solution lies in waste reduction rather than deploying new technology. This was echoed by respondents at Scania, who suggested that immediate deployment of new technology should not always be the initial solution because the optimal solutions are constantly evolving.

However, when performing production development by integrating digital technologies, it becomes evident for both companies that it triggers the tension of resource prioritisation between current operations and future SP needs. The production units are structured to enable control and resource efficiency, whereas the SP-team is organised to support an explorative approach, involving experimentation, variation and risk-taking (March, 1991). A similar dynamic can be observed in the relationship between LPS-teams and production, where the purpose of the LPS-team’s is to develop, maintain and disseminate Lean methods and tools. Respondents noted that individuals responsible for implementing solutions require time for this task, but that daily operations take precedence, resulting in less focus on digital transformation. Hence, allocating resources to production staff so that they can become part of the SP-team (or vice versa) is one key to become engage in the testing and development of new and future solutions.

In previous research, exploitation vs. exploration, is perhaps the most well-known example of the learning paradox (cf. March, 1991) and is also salient in the two cases. The importance of exploitation and exploration is expressed by one of the LPS managers at Scania: “Actually, not all new ideas are good ones. Most new ideas are bad ones, given the current context, but the context changes so nothing is static, and we still need to evolve”. The observations in Scania and Volvo CE highlight the importance of adopting new ideas in response to a changing context. Understanding the system, looking ahead and actively exploring new solution options appears to be a promising approach in addressing the dual need of exploitation and exploration. However, at both companies’ the respondents struggle with competence gaps in terms of not having a wide distribution of expertise or profound understanding of the digital technologies and their potential benefits for future LPS and vice versa. For example, limited technology competence amongst managers impacts the prioritisation of investments in new technology solutions, despite their deep knowledge of production and LP principles. Due to this, it was the experience of one of the respondents at Scania that it was difficult for senior management to answer questions in a qualified way regarding the prioritisation or relative importance of the technology focus.

To avoid competence gaps there are several initiatives to support continuous learning and competence development to enable managers and production operators to acquire new ambidextrous digital capabilities (cf. Secchi and Camuffo, 2019). Related to this is the tension between specialisation vs. generalisation emphasising the importance of integration of multiple learning perspectives and connecting production staff with different specialist functions and knowledge to extend the knowledge base in respect of the opportunities offered by digital technologies in manufacturing. The necessity to bridge the tension between specialisation vs. generalisation is evident in several statements by the respondents at both Scania and Volvo CE. For example, one of the LPS managers at Scania stated that: “We need people that serves as bridges and that sort of has a foot in both worlds. Right now, I believe that we are a bit too specialised” (see Table 3 for extended quote). At Volvo CE, the bridging of knowledge perspectives is partly addressed through creating dual affiliations of staff and creating small SP lab environments in which to meet across knowledge domains, which in turn connects the ways of managing the learning paradox to the belonging paradox.

The paradoxes of performing (current goals vs. future goals), organising (resource prioritisation between current operations vs. future SP needs) and learning (specialisation vs. generalisation) are all interconnected with the belonging paradox of integration vs. autonomy: The sense of being integrated and aligned with shared goals and ambitions for digitalisation and LPS, on the one hand and sense of autonomy and empowerment in improvement work to adopt new technologies or solutions, on the other hand.

Hence, the actions taken to manage the other paradoxes give rise to new paradoxical tensions related to identity and belonging. At Scania, it was observed that the introduction of new technical solutions affects work roles and identity in several ways. For instance, individuals experienced being disintegrated and a decrease in responsibility over things they previously had control over and some solutions were implemented without considering internal specialist functions or the actual operational needs, aptly illustrated by one of the respondents from operations at Scania: “We have the specialist functions. They are becoming increasingly frustrated. Why does nobody listen to us?” At Volvo CE, the learning paradox of specialisation vs. generalisation was managed by having SP-team members divide their time between working in the SP-team and performing various roles in the line organisation. Consequently, they spend part of their time focussing on the exploration of new digital technologies and seizing new opportunities aligned with the shared digitalisation strategy. For the remainder of their time, they returned to daily LPS operations (line organisation), a context that was heavily focussed on daily issues and thus had limited interest in new technologies, often showing resistance. This dual affiliation placed SP-team members in a crossfire of conflicting interests between their ongoing efforts towards the integration of SP goals and the autonomy and loyalty to their colleagues in the line organisation who partly resisted SP ideas. Hence, these dual affiliations result in dual loyalties that need to be balanced, also laying the groundwork for a belonging paradox. A critical issue for acceptance of the paradox then is to acquire a sense of ownership and relatedness to the digitalisation and improvement work by building trustful relationships across the organisation.

What becomes evident in the data analysis is that even though there is a recognition and acceptance of critical organisational paradoxes that need to be worked through (see Figure 1), there is at the same time a tendency of internal drift that amplify tensions and force imbalances of organisational paradoxes, i.e. disturbing the state of equilibrium, see Figure 2 (Weiser and Laamanen, 2022). Different kind of forces are constantly pushing the organisations towards imbalances and in the worst case they tilt the organisation and generates vicious cycles (Smith and Lewis, 2011) constraining the potential for integration. In the analysis, we discern four different imbalance archetypes where the risk of organisational tilt is high: Myopia, Dogma, Technical Monuments and Detached (Table 3).

Myopia: One of the most pressing issues reported in both companies that constrain the integration of digital technologies in LPS, is the scenario in which production becomes too short-sighted and gets stuck in the day-to-day production routine and becomes protective of the current ways of working, i.e. Myopia. A mechanism enabling Myopia is when the short-term goal of production is over-prioritised and generate an imbalance in the performing paradox, i.e. current goals become prioritised over future goals. A potential consequence of prioritising short-term goals is that this also amplifies tensions and triggers an imbalance of the organising paradox because it affects the resource prioritisation between current operations and the future needs of improvement work and digital technologies. When critical resources are not provided to support the opportunity to absorb new technical solutions into production and as a part of the LPS it could eventually constrain production innovation. The SP manager at Scania describes this as follows:

Another mechanism enabling drift towards Myopia is the situation in which production operators are consumed by their daily production routines and do not have the energy to look beyond their day-to-day work to carry out any future oriented improvement work, even though there be small windows of free time. Hence, daily production and disruptions in production affect the focus on integration of digital technologies because the daily disruptions are always prioritised, for instance to secure resource availability. In the case of Volvo CE, the biggest challenge in the implementation of digital Lean boards was to have a group that is working on the implementation every day, not just in one cell but in all cells.

Technical Monuments: A second critical issue is when the intention to integrate digital technologies becomes too technology-dominated, which generates an imbalance referred to as Technical Monuments. Technical Monuments’ was a phrase used by one of the respondents at Scania to describe the situation in which the introduction of new digital technology is imposed on production and, because of competence gaps, restricts the production organisation’s ability to absorb and utilise the principles of LP to improve.

The drift towards Technical Monuments occurs then as an imbalance of the organising paradox when technical solutions are imposed on production; solutions which are perceived as “good in themselves” by their proponents in the SP-team but which do not necessarily fulfil any need that has been identified by production:

At the same time, there is no “sense of pull” from production or LPS-team because of a competence gap amongst the users (operators) on how to maintain and improve the new machinery. Hence, the absence of proper coordination and alignment amongst stakeholder interests (Pozzi et al., 2023; Mokudai et al., 2021) renders an imbalance of the organising paradoxes, which as a consequence also triggers an imbalance of the belonging paradox as responsibilities changes in production and LPS-team; shifting responsibility for updating and maintaining the technical solution further away from the value adding process, making changes and improvements difficult and time consuming; circumstances which make the new solution inflexible and gradually turns it into a technical monument that suppress LPS-development.

Dogma: A third critical issue, Dogma, arises when the basic assumptions of LP and the LPS become axiomatic and not seen as principles that continuously needs to be adapted and as such becomes dogmatic. In the interviews, respondents describe a situation when LP principles are perceived by their proponents as a normative ideology and that “this is right in itself,” leading to a dogmatic view favouring the application of current ways of adopting Lean principles, which triggers an imbalance of the learning paradox because it suppresses the exploration of future orientations.

Furthermore, by overly focussing on the exploitation of current production through continuous improvements and elimination of waste, there is also a risk of missing out opportunities for exploration and disruptive change. In the case of Volvo CE, LPS-coaches can sometimes be focussed on “what and why are we going to change” and as such sometimes question the need for exploration.

The Dogma imbalance resemblances what Mokudai et al. (2021) refers to as the ”Lean trap” in their study of digital technologies as Lean augmentation. The Lean trap unfolds in a situation when production is strongly rooted in a LPS and this tends to make managers and production engineers only view the potential of digital technologies from the perspective of existing LP practices, and in so doing, they then risk being blind to the potentials of new technologies.

Detached: the data also includes indications of situations when senior management is perceived to not share the same pace of learning as the other stakeholders of the LPS-team, SP-team and production and as such becomes Detached. Hence, Detached is partly a consequence of not embracing the learning paradox in the sense that senior management are not up to date in exploring of future opportunities thus generating a competence gap.

The competence gap might then have spin-off effects because it triggers an imbalance of the performing paradox, as it impacts upon managers' opportunities or abilities to make decisions about new strategic initiatives that are grounded in the organisation.

In the previous two sections we have learnt about the critical organisational paradoxes embraced in the integration of digital technologies in a LPS and the organisational and human forces amplifying tensions and imbalances of organisational paradoxes and disturbing the state of equilibrium. In this section, we address RQ2 and focus on the micro-shifts of corrective actions (Weiser and Laamanen, 2022), what here is conceptualised as re-contextualising, which seeks to rebalance and conjoin the different poles of a paradox and as such cope with the tendencies of internal drift, whilst offering proper coordination and alignment between LP and SP activities (Pozzi et al., 2023; Mokudai et al., 2021). Based on the data of what contributes to acceptance and embracement of the identified paradoxes and other resolution measures taken in the companies to restore balance and drive change and transformation of manufacturing operations, four types of re-contextualising actions and strategies are identified: Empowering, Patching, Pacing and Leveraging (Table 3).

Empowering: learning paradoxical thinking. To avoid competence gaps within critical domains – such as digital technologies – and the risk of drift from the learning paradox into Dogma or Detached, there are in both companies several initiatives and support structures for empowering through continuous learning and competence development to enable managers, production engineers and operators to acquire new critical knowledge, e.g. digital capabilities. Confirming the findings of previous research (Dieste et al., 2022; Smith and Beretta, 2021), both companies emphasise the importance of investments in continuous learning as a key factor for ensuring that staff in production can re-contextualise and manage the complex interdependencies simultaneously and thus manage ambiguities so that they can make decisions for themselves and in the long run increase the possibilities of the pull effect of new technology to rebalance the organising paradox of pull vs. push. Therefore, empowering people is key to paradoxical thinking (Xiao et al., 2019).

Patching: connecting and converging knowledge elements. The SP-teams have toned down the “push” of new technology and instead emphasise the importance of demonstrating the potential of the digital technology. The shift is a re-contextualising through connecting and converging, i.e. patching, different knowledge perspectives, finding a fit between them. Patching can then be described as a rebalancing strategy for all four organisational paradoxes using different LP-practices (Johansson et al., 2022). One example from Scania is where personnel developed and tested a checklist to ensure that different knowledge perspectives and interests are taken into consideration when testing new digital technologies. Another example of activities supporting patching is seconding people between business units so that they gain experience from different parts of the organisation and better understand different stakeholders, e.g. production staff with different specialist functions working within the SP-team. As such, patching is also conducive to generating a sense of ownership and relatedness during integration work by building trust-based relationships across the organisations.

Pacing: synchronising time horizons. The different pace of changes and new initiatives initiated by the SP-team (long-term), continuous improvements in production (short-term) and the pace of day-to-day production also leads to tensions which act on other activities in the organisation. Thus, initiatives taken in one part of the organisation creates tensions due to the difference of time perspectives which at worst results in Myopia or Technical Monuments. Pacing becomes critical in re-contextualising and synchronising time perception between different stakeholders. Hence, by synchronising different time perceptions the nexus of current and future production needs and digital technology affordances can be identified and the receiver side of a new digital technology may alter the perception from a “sense of push” into a “sense of pull” through having new digital technologies available “just-in-time” to meet the needs being experienced.

Leveraging: utilising divergence for exploration. The importance of adopting new ideas in response to a changing context is a key factor in leveraging. By understanding the system, looking ahead and actively exploring new solution options, appears to be a promising approach when addressing the learning paradox of exploitation vs. exploration (cf. Secchi and Camuffo, 2019). In leveraging, the companies make use of tensions and the divergence in the organisation as a driver of exploration, which ironically, is partly the same as driving the imbalances of Dogma and Technical Monuments. However, by continuously highlighting tensions, the divergence in the organisation is explicated, which can then be used to re-contextualise and leverage transformation processes. Hence, leveraging is about managing the learning paradox of exploitation vs. exploration as transforming mechanism (March, 1991).

Previous studies have demonstrated the advantages of incorporating digital technologies into existing LP-management models (Buer et al., 2021; Tortorella et al., 2019), and further studies have utilised a paradox perspective when investigating the implementation of either I4.0 (e.g. Dieste et al., 2022) or LP (e.g. Maalouf and Gammelgaard, 2016), see Table 1. However, there is a gap in research examining paradoxes arising from the attempts to integrate LP and digital technologies (e.g. Holmemo and Korsen, 2023). Hence, this paper aims to advance the understanding of tensions and their paradoxical relationship and potential strategies for manage tensions when integrating digital technologies into existing LPSs. This study only highlights few contradictions between the aims of LP and digital technologies on a conceptual level. The respondents provide evidence of the potential complementary effects between LP and digital technologies, confirming the conclusion of Buer et al. (2021) that the simultaneous use of the two has a greater effect than the individual effects of each approach. However, our findings indicates that contradictions appear on a more detailed practical implementation level as each perspective generates different options and priorities, leading to paradoxical tensions. It is on this microfoundations level the paradoxical tensions need to be understood and managed (Maalouf and Gammelgaard, 2016; Dieste et al., 2022; Frank et al., 2019; Smith and Beretta, 2021; Erthal et al., 2021). More specifically, to gain answers to RQ1, the findings of this paper provide insights into the salient paradoxical tensions embraced in the integration of digital technologies in LPSs, by identifying different aspects of the performing, organising, learning and belonging paradoxes (Smith and Lewis, 2011). When addressing RQ2, the findings show the intricacies and relatedness between different paradoxes and their resolutions and, more specifically, how a resolution strategy adopted to manage one tension, might unintentionally generate new tensions or imbalances (Poole and Van de Ven, 1989). For example, the structural separation of LPS-team and SP-team is adopted to manage the performing paradox of competing but potentially synergetic approaches, on the one hand, which, however, disconnects and trigger an internal drift of the belonging paradox, on the other hand. This in turn calls for either re-contextualising actions to counteract the drift or the adopting of new resolution strategies, e.g. synthesis (Dieste et al., 2022) by organising cross-functional work across specialist functions such as the LPS- and SP-teams and production. In full accordance with the theoretical assumption made in the dissipative equilibrium model (Weiser and Laamanen, 2022), our findings showcase the tendency of internal drift because there are forces that constantly push the organisations towards imbalances, which in worst case tilts the organisation and generates vicious cycles constraining any integration of LP and digital technologies (Smith and Lewis, 2011). By identifying the micro-shifts and forces for the four imbalance archetypes Myopia, Technical Monuments, Dogma and Detached, we also highlight the sensitiveness and temporality (Poole and Van de Ven, 1989; Dooley and Van de Ven, 1999) of the state of equilibrium for all four paradoxes. Well in line with Smith and Lewis (2011) and Weiser and Laamanen (2022, our findings indicates that a dynamic equilibrium is not possibly to maintain without the micro-shifts of re-contextualising seeking to rebalance and conjoin the different poles of the paradoxes. This approach copes with tendencies of internal drift by enabling paradoxical thinking (Xiao et al., 2019) or ambidextrous capabilities (Secchi and Camuffo, 2019). Hence, resolutions and re-contextualising of paradoxes are always contextually sensitive and system dependent (Schad and Bansal, 2018).

Dieste et al. (2022) added the perspective of the dynamics and complexity in managing paradoxes in integrating digital technologies by suggesting four paradox resolution strategies: contextualisation, spatial separation, temporal separation and synthesis. Our contribution, based on micro-level analysis (Poole and Van de Ven, 1989) further adds to understanding of this complexity by also identifying the types of pulling apart forces and drift from equilibrium and the related integrating and rebalancing actions here framed as the re-contextualising actions of Empowering, Patching, Pacing and Leveraging. Hence, one key theoretical implication of our findings is that it adds micro-level understanding to previous research by empirically show the interplay between acceptance of paradoxes, pre-emptive balancing and re-contextualising as rebalancing mechanisms (Figure 3) and as such contributes with empirical evidence to the conceptual model of the dissipative equilibrium proposed by Weiser and Laamanen (2022). All in all, our findings shed new lights on how to understand rebalancing and transforming mechanisms to maintain dynamic equilibrium (Smith and Lewis, 2011; Weiser and Laamanen, 2022) and new evidence for the appropriateness of the paradox theory to uncover the complex intricacies and relatedness of the two perspectives.

In conclusion, our main contributions to the OM research and practice on LP vs. Industry 4.0 are three folded. First, the paper provides a fine-grained perspective on why integration sometimes “fails” and identifies the forces of internal drift as mechanisms of imbalances. By providing insights into critical paradoxes, our findings may help practitioners to understand the underlying reasons and its complexity, the ubiquitous risk of imbalances and create incentives to find a good balanced approach for SP. It may further help practitioners working on the integration of LP and SP to understand each other’s perspectives and contribute to create a common base for dialogue. Second, the paper provide detailed insights into how different management and resolution strategies are adopted, particularly by identifying re-contextualising actions as a key in rebalancing organisational paradoxes in favour of the integration of digital technologies in a LPS (see Figure 3). A critical issue in companies working on integrating LP and SP, is the constant need of rebalancing and corrective actions avoiding internal drift towards the extremes and to avoid the suppression of either side of organisational paradoxes. Hence, functional management strategies are all about avoiding getting stuck in the extremes. As such, this study has a robust anticipation power, helping managers/coaches responsible for driving LP and digitalisation initiatives in companies to find a balanced approach. Without this type of lens, contradictions and risk of internal drift cannot be anticipated. Third, considering that previous research has proven the adequacy of using paradox theory to investigate integration of LP, this study has also proven the usefulness of paradox theory in investigating the micro-level implications of integration of the two perspectives as part of production system development.

Finally, the two companies considered are well established in the LP-paradigm, whilst they have a shorter history of integrating digital technologies. To further elaborate and test the suggested paradoxes, tendencies to drift and management strategies for rebalancing and transforming dynamic equilibrium, future research would benefit from flipping the perspective and investigating companies that have their starting point in the SP paradigm and intent to integrate this with LP. Furthermore, applying the paradox perspective to SP transformation in SMEs would also be useful in providing a different setting where specialist roles and functions are more interwoven because of the size of the companies.

Funding: This work was supported by Excellence in Production Research (XPRES), a government-funded Strategic Research Area (SRA) within manufacturing engineering in Sweden.