Disruption handling by first-line managers in manufacturing: the situated active engineering of resilience Open Access

This study clarifies how first-line managers (FLMs) use situated action strategies to adapt and engineer resilience to handle severe disruptions in their manufacturing departments. Moreover, it examines three logically different types of disruptions and details the specific strategies required to resolve them in the studied companies.

Our study's primary theoretical contribution is the identification and analysis of three distinct disruption-specific logics (personnel-based, technical failure and complexity-induced). This framework moves beyond a monolithic view of disruptions to offer a novel, fine-grained conceptualization of how resilience is actively engineered. We advance the understanding of FLM work by showing that adaptability is not a generic skill, but a targeted response tailored to the specific logic of the disruption at hand, thus providing a new lens for analyzing resilience-in-action.

The study shows how manufacturing companies can strategically engineer resilience through their FLMs. By understanding the different logics of disruptions, organizations can move beyond one-size-fits-all emergency plans. They can develop targeted training and routines that prepare FLMs to deploy the right strategies (be it negotiated coordination, procedural escalation, or goal-oriented improvisation). The study also highlights the practical necessity for senior leadership to empower FLMs, allowing them to organize outside of normal hierarchies and rules to handle complex disruptions effectively.

The work of FLMs in manufacturing (in this paper, “frontline” is also used interchangeably with “first line”) is one of the most complex and demanding kinds of managerial work that can be found in manufacturing (Rabey, 2008; Tengblad, 2012). FLMs face diverse responsibilities and navigate intricate organizational structures (Townsend and Russell, 2013); the expectations on FLMs are multifaceted, widespread and complex. FLMs are considered to have a large impact on the performance of their work domain, and, as a group, to have a large impact on the company's performance and the ability to deliver according to volume, quality and time. FLMs are the managers who are responsible for transforming the strategies and goals of the company into stable performance in the front line of their part of the workshop (Slack and Lewis, 2017). In their role, FLMs collaborate with many different actors inside and sometimes outside the firm and they are affected by decisions made by many different expert functions and managers (Karltun et al., 2023).

Frontline production is characterized by variability and uncertainties, emanating from outside the company, usually as market-related, or from the inside, related to the uncertainties in the manufacturing process (Kara and Kayis, 2004). Front-line production can further be considered as a complex socio-technical system (STS), which is characterized by unanticipated variability (Saurin and Gonzalez, 2013). FLMs in manufacturing thus handle a set of common unanticipated occurrences, including machine breakdowns, delivery delays, material shortages, quality problems as well as absenteeism (Karltun et al., 2023). A central argument of this paper is the need for a clear conceptual distinction. We differentiate between the operational variability and unanticipated events that can be managed within the existing workflow, often through routine adjustments or adaptations. In contrast, we define disruptions as more severe events that cause a fundamental stoppage of production, demanding a shift in operational logic and a higher order of adaptive response.

Creating a planned stable output requires from FLMs a combination of decision-making, communication and leadership skills, paired with technical expertise and the ability to manage stress, supported by communication tools, knowledge management and monitoring systems. When empowered to contribute to continuous learning and improvement, FLMs contribute to resilient outcomes. This is enabled by their focus on day-to-day performance, leveraging their deep understanding of the local production system and interpreting data to anticipate and monitor, coordinating resources and communicating clearly to respond effectively and quickly as well as analyzing causes and suggesting improvements, contributing to organizational learning (Karltun et al., 2023).

However, the research on the concrete work of FLMs is scarce and as Townsend and Kellner (2015, p. 104) noted, “far too often the FLM has been a research finding rather than a research focus”. The aim of this paper is thus to investigate FLMs' action strategies to encounter disruptions in their daily work and how they thereby engineer resilience. By examining their daily flow of activities, this paper explores how the work of FLMs is adapted to coordinate work and mitigate the impact of disruptions as well as facilitate learning from them. Drawing upon intersecting relevant theoretical perspectives and empirical data, this study seeks to unveil how frontline management practices contribute to organizational resilience and maintain operational continuity.

This paper sheds light on the domain of frontline management in manufacturing, specifically focussing on how FLMs handle disruptions to maintain operational continuity.

The role of the FLM has been under development for a long time. From Taylor's (1911) suggestion of eight managerial specialized roles to more recent requests for a coaching role, where a lot of emphasis is put on enabling others (Townsend and Kellner, 2015). Key tasks for FLMs were found to be monitoring/controlling absence, handling discipline, managing the health and safety aspect of their staff, team development, introduction and training of their staff, planning/allocation of work, recruitment, appraising their staff, handling grievance and briefing their team (IRS, 2000) cited by Martins (2007). Townsend and Kellner (2015) also provide a comprehensive overview of the development of front-line manager work in general. In manufacturing and specifically in the companies studied in this paper, the influence of lean production and lean management was obvious. As notified by Poksinska et al. (2013) introduction of lean production changed managerial tasks to a more people-oriented work instead of process-oriented. The inclusion of leader standard work such as meeting structure, regular gemba walks including visual control, continuous improvement activities, two-way communication flow, coaching and training subordinates (Reynders et al., 2022) are important part of leadership work in lean. Delbridge and Lowe (1997) note that the daily work of FLMs is characterized by contradictory objectives but also complexity, high levels of social interaction with subordinates and delivery pressures.

At the frontline, unanticipated occurrences are frequent and based on the general variability and complexity of manufacturing systems. In their daily interaction with the system, the operators cannot be considered as constant; we differ from each other, and we do not perform equally every day. Moreover, the manufacturing system also differs in performance due to technical reasons, failures, etc (Guérin et al., 2007). To produce the intended result of the activity thus requires adaptations to the situated conditions all the time. The manufacturing systems studied here are considered complex STS, where the inherent complexity leads to variability. According to Saurin and Gonzalez (2013), these systems are characterized by the following features:

1. A large number of dynamically interacting elements

2. A wide diversity of elements

3. Unanticipated variability

4. Resilience, the ability of the system to adjust its functioning to uphold its performance under unexpected conditions (Hollnagel et al., 2011).

Other reasons for variability and requests for adaptations are market-related and manufacturing process-related, according to Kara and Kayis (2004). Here we concentrate on the requirements of flexibility in the workforce, where multi-skilled workers, numerical flexibility of workforce, part-time staff and overtime/idle time are suggested. All these aspects need to be handled by the FLM.

Root cause analysis (RCA) is a general concept used to deal with variability and enhance resilience within manufacturing organizations. Latino et al. (2019) highlight the importance of RCA in going beyond mere root cause investigation by uncovering not only physical causes of failures but also decision-making shortcomings and latent system deficiencies. By identifying and addressing underlying issues, RCA helps prevent the recurrence of failures and promotes a deeper understanding of how human actions and system factors interact to create risk. Effective RCA practices not only improve system resilience (Rooney and Heuvel, 2004) but also foster a “just culture” (Dekker, 2007) by promoting understanding of human decision-making in complex situations. While a deep investigation into the ultimate root causes of all external events (such as the societal factors behind absenteeism) is beyond the scope of this study, our focus is on how FLMs manage the consequences of such events within the production system. Thus, for our purposes, RCA is conceptualized as a managerial practice for understanding and mitigating operational failures and system deficiencies over which the organization has direct influence.

Woods' (2015) categorization of resilience perspectives distinguishes between rebound (the ability to bounce back from disruptions), robustness (the ability to withstand disruptions within known limits), graceful extensibility (the ability to extend capacity and adapt to surprises) and sustained adaptability (the ability to continuously adapt and learn over time).

To ensure conceptual clarity throughout this paper, we adopt the following definitions. Following Hollnagel (2018) and Woods (2015), we define resilience engineering as the overarching scientific paradigm that studies how complex socio-technical systems succeed under varying and unexpected conditions. In contrast, we use the term active engineering of resilience to denote the specific, situated actions and strategies employed by actors within the system (in our case, FLMs) to anticipate, monitor, respond to and learn from events. This distinction allows us to use resilience engineering as our theoretical lens while focussing our empirical analysis on the concrete work of engineering the resilience in practice. Dekker (2017) emphasizes the importance of leadership adaptation in fostering resilience, highlighting the need for leaders to adjust their approaches based on the specific situation. This study further draws upon Grote's (2019) studies to understand how FLMs adapt their leadership approaches and utilize various coordination mechanisms to maintain control while fostering adaptability in their teams. She highlights the importance of leaders being able to sense changes in demands, switch between different modes of operation and foster a culture that values both stability and flexibility. Grote (2019) further highlights the critical role of leadership in fostering the active engineering of resilience by enabling individuals and teams to adapt to complex demands. She argues that successful performance in dynamic environments relies on the ability to switch between different modes of operation, balancing stability and flexibility.

Analyzing the interplay between “safety by compliance” and “safety by resilience” (Grøtan, 2014, 2017) can shed light on how FLMs navigate the often-competing demands for stability and flexibility in their daily work. Grøtan argues that these two approaches are not mutually exclusive but rather coexist in a dynamic tension. While compliance with rules and procedures provides a necessary foundation for both safety and efficiency, the active engineering of resilience requires the ability to adapt, improvise and learn from unexpected events. This aligns with Hollnagel's (2018) Safety-II framework, which emphasizes that variability and adjustments are normal in everyday work. Variability and adjustments, therefore, can be seen as valuable opportunities for learning and adaptation.

Recent developments in resilience engineering have increasingly focused on quantitative assessment and measurement. For instance, Falegnami et al. (2025) have introduced models for measuring organizational resilience potential, while others have applied frameworks like the Resilience Assessment Grid (RAG) in sectors such as transport and healthcare to produce benchmarkable metrics (Klockner and Meredith, 2020; Safi et al., 2022). While this quantitative turn provides valuable tools for assessing an organization's adaptive capacity, it creates a corresponding need for rich qualitative research that explores the underlying processes and situated practices of how this capacity is enacted in real-time. Our study addresses this gap by focussing on the micro-level strategies FLMs employ to navigate disruptions, thereby complementing the macro-level assessment literature with a grounded understanding of resilience-in-action.

The research adopts a collaborative multiple-case study approach (Yin, 2018) in close cooperation with two manufacturing companies regarding FLMs' action strategies in managing their daily work. An interactive design, such as engaged scholarship, was used as a common platform to bridge the gaps between theory and practice (Ellström et al., 2020; Van de Ven, 2007). The research was carried out together by three project groups: one at each company (7–8 participants) and one with the five researchers. A detailed description of the research approach is published as a research protocol (Karltun et al., 2021).

Two Swedish manufacturing companies were purposively selected to take part in the study. The criteria for their selection were their strong reputations as successful, long-term implementers of lean manufacturing principles and their willingness to engage in a multi-year collaborative research project, making them ideal sites for studying advanced managerial practices in resilient organizations. Both companies were successful within their specific market segments and enjoyed strong positive reputations, being models for many other companies. They had successfully implemented many of the core principles of lean (Liker, 2004) within their operations for many years.

The lighting company specialized in the development, manufacturing and marketing of professional lighting solutions for public spaces. The participating factory site involved in the project employed approximately 600 people and had 8 FLMs, each overseeing between 27 and 48 employees. The production character varied, ranging from standard, simple luminaires assembled on production lines to large, single architect-designed high-end luminaires as well as specialized gas panels tailored for hospital use. New products were introduced frequently, and the company had recently undergone a major technological transition to exclusively use LED (Light Emitting Diode) lighting.

The truck company manufactured truck cabins for the European market. The factory site involved in the project employed around 2,700 people and had recently made significant investments in automation. Approximately 100 FLMs worked at the site, each overseeing between 12 and 40 frontline workers. The production process was characterized by large-scale manufacturing of customized cabins, each one tailored to specific requirements, yet all produced within the same standardized production flow. The company had a strong reputation for working integrated with quality improvement and lean, being one of the companies mentioned in Bergman et al.(2024).

The selection of participating FLMs was conducted as a collaborative process between the research team and senior management at each company. The criteria for inclusion were twofold: first, to ensure representation across a wide spectrum of managerial experience (from 1.5 to 22 years as an FLM) and second, to select individuals who demonstrated a strong, self-professed interest in reflective practice and the development of their own work. This second criterion was essential for the collaborative, interactive nature of our research design. While the initial pool of candidates was nominated by the companies, all participants were fully briefed on the study's aims and methods before giving their informed consent, confirming their suitability and willingness to engage in the in-depth data collection required. As the overarching aim of the research collaboration was the development of FLM work, both experienced and less experienced FLMs were chosen. All, however, showed an active personal interest in developing their own work (see Table 1).

Data was collected between September 2019 and March 2022. In examining how FLMs' work was performed, attention was directed toward issues related to the individual as well as the inherent variabilities of the organizational work settings. The work was decomposed into discrete tasks, according to the expected results of work and into work activities considered as the way the tasks were accomplished (Guérin et al., 2007). This was achieved through semi-structured interviews and workshops, combining the FLMs' personal narratives with data gathered by the researchers.

The semi-structured interviews were conducted with FLMs, support functions and other members of the organization. This included frontline staff, service personnel like quality officers, maintenance and technical staff, HR (Human Resources) representatives, schedulers, as well as higher-level positions (see Table 2). The interview guide covered several themes inspired by the RAG procedures (Hollnagel, 2018), focussing on resilience potentials, along with general questions regarding the organizational structure of the company. The interview guide consistently addressed respondents' daily activities, role perceptions, interactions, decision-making and experience in managing diverse situations and unanticipated variability related to organizational requirements.

Gaining insight into how the companies operated from an activity-based perspective involved creating a rich picture of the context surrounding the FLMs' work activities, particularly how these activities were influenced by organizational and technical conditions. This understanding was developed through a series of iterative workshops that included participation from FLMs, HR officers and production managers (see Table 2).

A total of 23 workshops were held both within each company individually and jointly between the two. Eight of the workshops were dedicated to exploring and evaluating FLMs' work practices through critical reflection and benchmarking. The following fifteen workshops concentrated on activities aimed at improvement and development. The workshop groups were formed either within each company or as joint sessions centred on particular themes.

All interviews were audio recorded, and workshops were video recorded. In addition to interviews and workshops, document analysis was conducted, including formal requirements, role descriptions, routine documentation and process charts.

In the first phase, all interview transcripts were analyzed for each participant using the themes outlined in the interview guide template. Specific examples of FLMs' actions in response to disruptions within each company were then categorized according to the four resilience potentials defined in Hollnagel's framework (2011, 2018). Additional attention was given to contextual factors that either hindered or supported the development of operational resilience strategies aimed at managing various disruptions in each company. The documentation of this analysis served as the foundation for eight problematization and benchmarking workshops and later 15 developmental workshops with subgroups (see Table 2). Dialogues were conducted with each company followed by joint discussions involving both organizations.

In the second phase, in-depth analysis of data from the eight problematizing and benchmarking workshops was undertaken. The participating FLMs reconstructed (Gadamer et al., 2004) several cases during the workshops, where they made flow charts to visualize various disruptions and described in each case the handling of the disruption in detail. All cases were followed by an interactive analysis, with researchers and participating FLMs together. The cases were broken down into different aspects of how each case was handled, related to the degree of complexity, policies and organizational requirements. Three of these unexpected disruption experiences (Figure 1, Figure 2 and Figure 3) are focused on in this paper.

Our analytical approach was conducted in two distinct phases, reflecting the multi-level nature of the phenomenon under study. The first phase involved a thematic analysis of the 32 semi-structured interview transcripts. This provided a rich understanding of the subjective, individual experiences of FLMs and support staff, capturing their personal interpretations of daily events, operational variability and organizational context. The second, crucial phase of analysis was integrated within the eight problematizing and benchmarking workshops. These workshops were designed to move from individual accounts to a collective, intersubjective reconstruction of major disruptions. In these sessions, participants collaboratively mapped system-level event flows, engaged in a joint analysis of causal chains and verified interpretations in real time with the research team. This second phase generated a distinct, system-level dataset that captured the coordinated and interactive nature of disruption handling.

Given this two-tiered data structure, we argue that a synthesized narrative is the most appropriate and rigorous method for presenting our findings. A fragmented presentation of individual interview quotes would fail to capture the systemic, coordinated and intersubjective understanding of disruptions that emerged from the workshops. The narrative approach, grounded in the principles of narrative inquiry which seek to understand experience as lived and told (Bruner, 1988; Clandinin and Connelly, 2000), is therefore not a simplification but a strategic analytical choice. It allows us to integrate the subjective “lived experience” from the interviews with the objective “system-level process” from the workshops into a coherent, holistic representation of the active engineering of resilience as it unfolds in time and across organizational boundaries.

The findings section is built by a narrative and a primary analysis of the narrative.

This narrative, a composite of multiple FLM accounts, describes a fictitious working day of an FLM at the production line and it focuses on the regular structures and meetings that occur every day, but also includes planned events that occur irregularly and three totally unexpected events. All details are based on real situations extracted from multiple FLM accounts gathered in the data.

07.00 The working day begins with a short meeting with the team leaders at the control board. We check the production situation, what is to be achieved during the day, staffing, backlog from yesterday, work environment and safety. The operational HR officer usually participates.

07.30 The workshop teams and production support functions report on what known problems that need to be handled. These are handed over to relevant support functions, who are expected to solve the problems during the day.

08.00 Meeting with all FLMs, where we report on meetings with team leaders and outcomes per FLM area. The second line manager leads the meeting. Any unanticipated events are reported and allocated to responsible managers. When that meeting is over, you have a good idea of what will happen during the day. Today, the meeting needs to be prolonged to handle a staffing situation as there is an unexpected lack of workforce. To keep production running during the day, the staffing between the departments needs to be adjusted. To maintain enough competence in each department, a negotiation takes place where staff are temporarily moved between departments and some production is replanned. A lot of issues need to be considered to make this right and during the meeting, contacts are also taken by phone and through the IT (Information Technology) system (Figure 1).

09.00 Meeting with the operative HR manager, union representative and occupational health officer in order to follow up on a rehab plan. The mobile phone rings, but I choose not to answer; the calling team leader can be contacted later. New call from the same team leader, this time I answer as an immediate repetition signal urgency. The team leader informs me about an acute critical situation where a ball valve has fallen from its position under the ceiling over a production line. An operator was close to being hit by the valve and a lot of coolant leaked and sprayed out. I immediately leave the meeting and hurry down to the situation. Production has stopped and the team leader has blocked off the area. She also gathered the work group to calm down and inform them about the situation. Moreover, she had already contacted maintenance technicians, who in turn contacted the real estate department. At the same time, I escalate the event to the second line manager, who escalates further to the third line manager. Both are quickly in place. Action at stops starts (a method for handling production stops). Decisions are made at the third line manager level, where necessary resources are available, to build a protective scaffold to stop the leak, make a repair and check all similar valves. Information goes back and forth between me, maintenance, team leaders and concerned managers. I inform other FLMs about the stop and that we have a temporary staffing surplus. Time is booked for the “learn from stop” meeting, Figure 2.

10.30 Weekly meeting with all FLMs and the second line manager. Reporting on the last day's performance and information on coming product projects. I raise the issue of assembly problems caused by new product projects. It is an ongoing problem with us and generally in the workshop. I feel that I am too little involved to develop a better understanding of this.

12.00 Lunch with other FLMs. It is a good time for us to solve problems, find ideas for different solutions. The FLM in another department talks about an event a few weeks ago about a problem with the assembly of a signal switch. It did not fit the instrument panel, even though the drawing showed that it would work. The FLM informed the scheduling department and contacted the design department, design contacted the marketing department and marketing contacted the two different suppliers that had been specified by the end customer for the product concerned. It turned out that the two suppliers did not talk to each other as they were direct competitors. Marketing then went back to the building contractor, who was the former customer and asked if they could contact the end customer to inform them that the specification in the contract was incorrect. The problem was tossed back and forth between our marketing department, customer contractor and end customer, says the FLM. With us, production stood still. We had to start another production as fast as we could. Meanwhile, the problem was moved back to purchasing, which in turn announced scheduling for when we could count on getting the right components. The test equipment needed to be fixed to completely test the function, and production engineering and purchasing were mixed in. The FLM chased all instances to get the product to the customer on time. It was a bit messy, many messages, decisions that had to be made, says the FLM. At the same time, the marketing would keep in touch with the contract customer and the end customer. In this case, it took weeks to get information and decisions from the end customer (Figure 3).

After lunch, I run a new check round in production. Receives an email from the HR manager stating that it is time to perform employee dialogues and that a meeting on new documentation routines will be held within a few weeks. Have a quick cup of coffee, talk to the production (third line) manager about the incident with the ball valve. A special group has been appointed to investigate the cause of the incident after the “learn from stop”-meeting.

14.30 I start planning for the evening shift. Before that, the team leaders update me on any evening needs of staff. Planning is based on the number of people and priorities. A search for remaining orders today is made, as this can contribute to prioritization. After the planning has been completed, this is emailed to the relevant work groups.

15.00 Myself and the team leaders make a situation assessment.

15.45 We run the outcome report and shift handover at the control board.

The days at work are intense; not two days are the same. I am very dependent on my team leaders, trust that the team leaders can solve problems, prioritize and re-plan. I have good use of information and communication systems (ICT) that help me to keep track of the production results in my department and to communicate with others, plan the work in the near future and other things that affect the department. In the computer, I have information about the current workload and for coming days, supply balances, financial outcomes and expected outcomes, salaries and other personnel-related information. The mobile phone is my main tool where I call, can send SMS (Short Message Service), email, but also use WhatsApp and notification features to reach everyone involved quickly.

We examine the narrative above using the model depicted in Figure 4. We can discern three distinct levels of work and the corresponding requirements on adaptability in the action strategies employed. By this, we provide an empirically grounded understanding of FLMs' active role in the engineering of resilience and extending adaptive capacity.

This level may be viewed as where FLMs' rhythm of daily management plays out. It encompasses the routine aspects of FLM work, characterized by a high degree of control and low complexity. It includes activities such as daily meetings with team leaders, communication with support functions and participation in recurring meetings with other managers. What needs to be done and how to do it is generally well-known and established within the organization's routines and procedures. At this level of adaptability, we find the routine variability and minor events that can be handled within the normal workflow, thus without any production downtime.

This level involves managing situations that, while not necessarily routine, are known in advance and allow for preparation. This includes scheduled meetings, training sessions and anticipated events like “Employee dialogues” and the “Rehab meeting” depicted in Figure 4. While these events may require FLMs to adjust their regular schedules and reallocate resources, established procedures and protocols can be followed. This level of adaptability can embrace more significant unanticipated events that can be managed without necessitating production downtime, but may require replanned schedules.

This level is achieved when encountering unexpected challenges and surprises that disrupt the normal workflow. It represents situations where the manufacturing system is fundamentally challenged by unforeseen events that fall outside planned routines and procedures. The consequences of these situations are true disruptions to production; it stops and cannot be started without involving several managerial levels and departments. Complexity is high, and control is low, requiring FLMs to improvise, adapt and make decisions with limited information.

The incident involving too high workload (Figure 1) highlights the FLM's role in maintaining operational stability. This situation required the FLM to rapidly assess the situation together with colleagues, engage in negotiation and communication, make decisions under uncertainty and activate organizational support mechanisms. The main target for these negotiations is how to optimize the resources available, which requires the involved FLMs to know at a detailed level what tasks can be performed by the present subordinates. This detailed knowledge, combined with team leaders' familiarity, makes it possible to share or distribute the available staff.

The incident with the fallen ball valve is a typical technical breakdown situation where urgency and protection of employees are important. Production at the affected department was immediately disrupted, and all available resources were redirected to manage the situation. The example illustrates a trained sequence of actions where essential people are involved directly in a task force to deal with the problem. Formal necessary staff and financial resources are acquired by direct escalation to the appropriate management level, following implemented routines for this. As soon as there is evidence that the problem is solved, beyond the immediate response, a follow-up meeting involving the FLM, affected team and relevant departments takes place. This highlights the value of post-incident learning, which allowed for a thorough analysis of the valve failure, identification of contributing factors and development of corrective actions to prevent recurrence.

The signal switch incident exemplifies how complex systems can produce unexpected outcomes. The FLM acted as a self-appointed project manager to solve the root problem, the incompatible signal switch. As can be understood from the description of the case, the work in the assembly department was entirely disrupted. All goods had to be removed from the assembly area; other products had to be started for assembly in parallel to trying to find a solution for the delivery of the signal switch assembly. The FLM approached several other departments and chased them daily or frequently during several weeks to get the problem solved as fast as possible. In doing so, he side-stepped the organizational hierarchy and took a more process and goal-oriented stance.

The work of the FLMs in this study can be divided into three levels of adaptability demands. Levels 1 and 2 concern the management of the operational variability and unanticipated events of daily work. In contrast, Level 3 involves the severe disruptions resulting in production stoppages, which form the core of our analysis. The request for immediate actions is paramount but diverse; the narrative illustrates three very different forms of disruptions that follow different logics and must be met by fundamentally different action strategies. In the first example, negotiation, communication, collaboration and developing multiskilled personnel are key aspects. In the second example, a trained task force organizing, including established escalation routines, are important. The third example shows a very distinct movement from following established rules within the organization to goal-oriented work.

This study provides a deepened understanding of how FLMs in manufacturing actively contribute to the emergence of organizational resilience by effectively handling variability and disruptions. Our findings illustrate three levels of adaptability, each corresponding to varying demands placed on FLMs depending on the effort required and the familiarity of the activity (see Figure 4). These findings align with Hollnagel's (2018) argument that resilient performance in complex socio-technical systems arises from the ability to adapt to changing conditions and unexpected events.

At Level 1 we find ordinary work, which can be related to leader standard work (Mann, 2015). On a structural level, there is also a high correspondence between work as imagined (WAI) and work as done (WAD) (Soliman and Saurin, 2022). At Level 2 we find work activities that are scheduled in advance but are not equally routinized and thus require a higher level of adaptability. At Level 3 we find three different acute, non-planned disruptions of production processes.

The first case is due to a large shortage of personnel. The logic of solving the disruption in this case is consequently to plan for multiskilling over department borders and to coordinate with closely situated departments for a common best solution. The criteria presented by Cook and Long (2021, p. 3) can serve as a checklist on why this can work. Most important are multiskilled workers, the availability of these and that other departments don't suffer from lending these out. Multiskilling to create flexibility and to mitigate high specialization of labour was the fourth principle of sociotechnical design published by Cherns (1976); moreover, it is an important intention in lean (Liker, 2004). The logic of solving the disruption in this case is consequently to plan for multiskilling over department borders and to coordinate with closely situated departments for a common best solution.

The second case was due to a technical failure (a valve falling down and collapsing the coolant flow for the machinery). This immediately started a safety process to protect the operators, an escalation chain to receive management involvement and support and created a task force to solve the short-term problem as well as hindering it from happening again. The problem was solved very rapidly due to well-trained personnel, established escalation routines and established work procedures for handling technical breakdowns. Interestingly, the task force also had a follow-up meeting to identify what to learn from the case. As emphasized by Ito et al. (2022), improved RCA and knowledge sharing are vital for engineering enhanced resilience in production systems. The logic in this case is thus well-trained procedures for handling technical breakdowns and ascertained learning procedures.

The third case bears clear characteristics of complexity failure. All involved had done their work according to normal routines. Still, the disruption occurred during assembly, and the product could not be delivered. After solving the immediate stoppage, the FLM got outside his normal role and started to act as a project manager. By doing so, he no longer followed established roles and procedures but instead acted goal-oriented to find a solution, moving from countering pathogenesis to facilitating salutogenesis (see Figure 5; Grøtan, 2014, 2017). Specifically, by acting as a self-appointed project manager and bypassing the formal hierarchy, the FLM moved from a state of “safety by compliance” to one of “safety by resilience,” demonstrating the adaptive, goal-oriented practice the model describes.

Figure 5 elaborates on the dynamics between compliance and engineered resilience. It depicts a continuum where organizations navigate between error prevention (compliance) and active adaptation to surprise (resilience). The overlapping circles and “stop criteria” illustrate the need to balance these approaches, acknowledging that rigid adherence to rules can hinder adaptability while excessive reliance on improvisation can compromise safety. This model emphasizes the importance of organizational learning and “salutogenesis” (fostering conditions that support adaptability and continuous improvement in complex socio-technical systems). It suggests that organizations can shift towards a more resilient mode of operation by embracing adaptive practices and learning from disruptions.

All cases at Level 3 require high levels of adaptability. The examples also illustrate the importance of switching between balancing stability and flexibility and the critical role of leadership in fostering the active engineering of resilience (Grote, 2019). It also demonstrates that even with well-performing lean systems, WAI and WAD exist in parallel (Soliman and Saurin, 2022). The findings also demonstrate the importance of coordination, the fifth aspect of resilience introduced in recent work (Karltun et al., 2023). This aspect was particularly salient in resolving Level 3 disruptions.

This study's primary contribution to the field of resilience engineering and manufacturing management is the identification and conceptualization of disruption-specific logics. Our findings demonstrate that not all disruptions are equivalent; they are triggered by different underlying causes (personnel shortages, technical breakdowns and complexity failures) and consequently demand fundamentally different adaptive responses from first-line managers (FLMs). By moving beyond a monolithic view of disruptions, we provide a novel, empirically grounded framework for understanding how resilience is actively and situationally engineered on the front line.

We situate these logics within three levels of required adaptability. Level 1 and 2 events, representing the baseline of operational variability, are managed through routine and semi-routine actions closely aligned with leader standard work (Mann, 2015), where work-as-imagined and work-as-done are in high correspondence. It is at Level 3 where true “disruptions” occur, fundamentally stopping production and demanding a shift in managerial strategy. Here, the distinct logics become paramount. We found that:

1. Personnel-based disruptions (e.g. high absenteeism) trigger a logic of negotiated coordination, a response that relies on pre-existing social capital, trust between managers and the planned flexibility afforded by multiskilled workers. This approach aligns with sociotechnical design principles (Cherns, 1976) and demonstrates the critical importance of adaptive capacity sharing to manage resource constraints (Cook and Long, 2021).

2. Technical failure disruptions (e.g. the fallen valve) trigger a logic of procedural escalation, a response characterized by well-trained routines, rapid task-force organization and clear hierarchical communication channels for resource allocation. The effectiveness of this logic is enhanced by robust post-incident learning, where improved RCA and knowledge sharing become vital for engineering future resilience (Ito et al., 2022).

3. Complexity-induced disruptions (e.g. the signal switch failure) trigger a logic of goal-oriented improvisation, a response that requires FLMs to transcend their formal roles and hierarchies, acting as cross-functional project managers and shifting from a compliance mindset to one of salutogenesis (Grøtan, 2014, 2017).

This typology of adaptive responses represents a significant step forward in understanding the concrete work of FLMs as key agents of resilience. It reinforces the importance of coordination as a core resilience capability (Karltun et al., 2023) and provides a textured illustration of the dynamic interplay between work-as-imagined and work-as-done in complex systems (Soliman and Saurin, 2022).

For practice, our findings suggest that organizations can enhance resilience not by creating a single, universal response plan, but by preparing managers to recognize and deploy different logics. This involves investing in multiskilling for negotiated coordination, training for procedural escalation and, crucially, empowering FLMs with the autonomy and senior support needed for goal-oriented improvisation. This approach aligns with Woods' (2018) concept of graceful extensibility and acknowledges that in the most complex disruptions, sustainable resilience may require managers to break established rules to achieve vital organizational goals.

This study has some limitations. The use of case study research offers a rich and engaging portrayal of FLM experiences, but the focus on two specific companies and their industry settings limits the generalizability of the findings. Future research could explore the engineering of resilience by FLMs in different industry settings and with varying levels of technological advancement. Additionally, investigating the impact of leadership styles and national context on how FLMs actively engineer resilience could provide further valuable insights and contribute to a more comprehensive understanding of the phenomenon.

While this study's case-based approach provides rich, contextualized insights, its findings are specific to two companies. Our framework of disruption-specific logics, however, provides a direct roadmap for future research. More qualitative research could identify further underlying logics and examine how our findings appear in other complex fields, such as healthcare or logistics. Quantitative studies could test the prevalence and impact of these underlying logics across a larger sample of firms. Such research would build directly on our contribution, further developing a nuanced understanding of how resilience is actively engineered in practice.

Johan Karltun: Writing – review and editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Conceptualization, Finance acquisition. Karin Havemose: Writing – original draft, review and editing, Visualization, Formal analysis, Investigation, Validation. Anette Karltun: Conceptualization, Methodology, Investigation, Visualization, Formal analysis, Validation, Writing – review and editing, Finance acquisition. Sofia Kjellström: Conceptualization, Methodology, Investigation, Validation, Writing – review and editing, Finance acquisition. Denis Alves Coelho: Writing – review and editing, Formal analysis, Conceptualization, Validation.

Declaration of generative AI and AI-assisted technologies in the writing process: During the preparation of this work the authors used Grammarly in order to improve the syntax in parts of the paper. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

The authors thank the participating companies and their personnel who actively contributed to this research study.