Managing business process reengineering within the framework of robotic process automation

This research is designed to examine how has traditional Business Process Reengineering activities changed with the implementation of Robotic Process Automation. A multiple case study of 3 large metallurgical manufacturing organizations based on semi-structured interviews with executives, BPR projects leaders and project teams' participants explores how process reengineering activities are rethought due to RPA technology deployment. The research provides the reimagined RPA-centered framework that explains the sequence of stages and activities performed on each stage. The framework shows that RPA implementation requires a novel approach to BPR from project leaders and executives: robotized processes are consolidated in one center; multifunctional teams are composed, and BPR stages are redesigned. This approach facilitates in frictionless RPA implementation. Understanding and using the new BPR framework is essential for smooth and successful RPA implementation and, in turn, high performance of the company. Project leaders' tacit knowledge that unsurfaced during semi-structured interviews provided feasible managerial implications for further RPA adherents from a similar business area.

Keywords: Robotic Process Automation, Business Process Reengineering, BPR framework, software bots, steelmaking industry, case study, Russia

Business Process Reengineering (BPR) as a concept has been in the business scene for more than 25 years and has been comprehensively researched by the scientific community. It is defined as a mixture of management guidelines, project management and industrial engineering techniques, modeling, design, prototyping and analysis tools that aims to thoroughly inspect current business processes in order for them to be radically redesigned for strong performance improvements (Love, Gunasekaran 1997; Gartner). BPR is named as “one of the most powerful” instruments that can stimulate customer satisfaction and boost firm performance (Mansar & Reijers, 2005). Most companies have recognized the benefits that could potentially be achieved by deploying BPR, including reduced costs and cycle time, higher competitiveness etc., and either outsource this activity to consultancy services or establish BPR teams in-house. Typically, BPR teams rebuild old processes from scratch by using numerous software solutions that have emerged in the recent years, and decide what activities to reduce, automate or eliminate completely and how to effectively allocate resources. Particularly automation can be applied to processes in different ways. Robotic Process Automation (RPA), as a distinctive area of automation, appears to be a relatively new tool or set of tools applied by companies to automate repetitive and standardized tasks so that they are performed by software robots (bots) or artificial intelligence instead of humans. Bots allow to recover hours of work, allocate employees to more creative projects that require complex decision-making. For a task to be automated by bots, it should be well-structured and rule-based, because bots are programmed follow protocols precisely (Deloitte, 2017). That is why RPA is mostly deployed in banking, insurance or telecommunication industries, since they possess process-aware information systems (Syed, et al. 2020; Gartner, 2018). Recently, there has been consistent growth in adoption of RPA solutions and by 2022, investments into RPA software will have amounted to $2,4 billion (Gartner, 2018). Implementation of software bots into current processes requires BPR activities and analytical approach that make sure right tasks are automated, it is completed as smoothly as possible, RPA is well integrated into an organizational framework and a desired economic effect is achieved. Without BPR efforts, deploying RPA might waste key firm resources such as people, technologies and processes, thus the full benefit potential of RPA will not be realized (Deloitte, 2017; Syed et al. 2020). Even though RPA deployment has considerably expanded over the past few years, and some consultancy companies like Deloitte readjusted their BPR frameworks in order to fit in considerations for robotics, it is not yet clear how in-house BPR teams and activities adjust and alter to seamlessly deploy RPA.

Although there has been significant number of papers written on BPR and RPA as separate research streams, the BPR framework in conjunction with RPA tools has not been well examined. Because of that, BPR within the framework of RPA is deemed a fruitful research topic. A recent research of RPA-related literature revealed a few gaps that need to be addressed. Firstly, there are no clear guidelines on how to realize the benefits of RPA solutions as they vary from industry to industry. Moreover, this subject lacks a framework of critical success factors and their implications, which should be viewed in different organizational contexts where RPA takes place (Syed et al., 2020). This paper attempts to address these gaps, by considering RPA through BPR activities. This research explores how has traditional BPR framework changed with increased implementations of RPA and design a new BPR framework that is RPA-centered. In particular, this research aims to discover what are the steps for BPR teams to undertake and whether new BPR teams are needed in a company with introduction of RPA. Moreover, how can RPA tools be incorporated into BPR framework itself and what are the critical success factors of such BPR projects. Therefore, the purpose of the research is both descriptive and explanatory. It is designed not only to overview how BPR framework changes with introduction of RPA tools, but also to obtain a better understanding of critical success or failure factors of BPR automation projects.

The research objectives are achieved by conducting a multiple case study of 3 large Russian steel manufacturing companies and its in-house BPR teams. The choice of the multiple case study strategy is justified by two reasons. Firstly, the companies chosen for the case study possess the features essential for the research: in-house BPR teams; both on-going and completed BPR projects; processes robotized via software bots (RPA). The second reason is that multiple case study has the potential to obtain a deeper understanding (Gustafsson, 2017) and provide useful insight on BPR activities in regards to RPA implementations in the steel manufacturing industry as the enterprises are currently leading in high-quality steel production on both Russian and the international markets (Deloitte, 2017). This cross-sectional research examines RPA-centered BPR projects of these companies at a specific point in time by analyzing internal documents that govern these projects and conducting semi-structured interviews with executives and RPA developers to extract both explicit and tacit knowledge on the topic. Interviews are recorded and transcribed, qualitative content analysis is then applied to examine transcripts, find connections between responses and draw conclusions. Based on the analysis, the reimagined RPA-centered BPR framework is designed and compared to the theoretical foundation. In order to further explore what is essential for RPA-oriented BPR projects to be performed successfully, positively contribute to achieving organizational goals and uncover the full potential of this technology, critical success factors are developed as the result of qualitative content analysis. Critical success factors are identified in alignment with existent research of the topic and specificities of the industry.

While more companies consider investing into RPA tools, managers lack practical knowledge on how to take them into consideration within the framework of already existing BPR activities and teams. Not only this paper has revealed a new BPR framework that can potentially be used by other similar businesses that opt for RPA solutions, but it has also provided concrete managerial guidelines for automation centered BPR projects to succeed and realize the full benefit potential of RPA. Thus, this research provides useful insight obtained from the real experience of BPR teams at 3 of the largest steelmaking companies in Russia and can be further used as part of benchmarking analysis for companies of similar industries and structures that opt for RPA solutions and want to employ in-house BPR teams for implementation. The study also provides an original contribution to existing literature by giving a comprehensive overview on how BPR and RPA function together and relate to each other in iron and steel manufacturing enterprise context.

1. Literature Review

1.1 Traditional framework of business process reengineering

Since this research is focused on how traditional BPR framework changes with introduction of RPA technologies, it is firstly required to identify what is considered a “traditional” framework of BPR. Kettinger et al. (1997) provided a reliable generalized BPR project Stage-Activity framework based on surveys and studies of a few leading reengineering consulting firms. The framework consists of 6 main stages (Envision, Initiate, Diagnose, Redesign, Construct, Evaluate), and each of them contains several activities. Envision stage is characterized by composing a task force that reviews business processes for potential improvement in accordance with company's strategy. At the Initiate stage, the actual BPR team is created, goals and plans are set via benchmarking or cost benefit analysis. Diagnose stage implies a thorough analysis of existent processes to identify areas for improvement, problems and activities that do not add value. During Redesign, brainstorming is used to create a new process design that is then documented and prototyped. At the Reconstruct stage, new process is implemented with its new IT systems, human resource roles and responsibilities; team works to ensure smooth transition. Finally, during Evaluation, the new process is monitored to ensure its consistency with set goals, company's strategy and structure. Even though this framework described most of the possible BPR stages, it may still vary and be altered with time due to the changing nature of BPR (Kettinger et al. 1997). Extensive literature research conducted by Motwani et al. (1998) yielded another six-phase comprehensive BPR framework that consists of Understanding, Initiating, Programming, Transforming, Implementing, Evaluating stages. When combined and transformed, these frameworks might represent a generalized traditional BPR framework that will be used as theoretical basis for this research (Figure 1).

In regards to BPR modelling and analysis tools and techniques, Gunasekaran and Kobu (2002) examined a scope of literature written on the matter in order to identify the most common BPR tools and techniques used in various fields of application and formulate guidelines for their selection based on the nature of BPR and corresponding areas. The framework for choosing appropriate tools was proposed in a form of checklist. For instance, the authors suggested using Petri-Net Models to understand the business process system as they provide visual forms that are easy to comprehend.

Figure 1 - Traditional BPR framework (transformed from Kettinger et al. 1997; Motwani et al. 1998)

For improving value-creating activities, it was suggested to use such tools as activity-based analysis, work-flow models and flow charts. Overall, a variety of different tools and techniques can be used at different stages of BPR from initiation to post-implementation: core process analysis, benchmarking, computer-aided software engineering, cognitive mapping, role activity diagramming, etc (Kettinger et al. 1997).

1.2 Business process reengineering critical success factors

Just as with any other activity, some issues may inevitably arise during BPR tools and techniques application. Fasna and Gunatilake (2019) reviewed a scope of articles on the matter and identified more than 60 possible issues that might occur on different stages of BPR: pre-implementation, implementation and post-implementation. Via the case study it was found that during the first BPR phase the most common problem is lack of top managerial engagement and support of the process. During the actual implementation BPR team members may face the challenge of misalignment between team responsibilities and functional responsibilities, while the last phases' most common struggle is measuring the overall project's performance. Issues that occur along the way of realizing BPR projects and the ways that teams overcome them might have substantial impact on the outcome.

Some researchers have investigated unsuccessful BPR projects and analyzed reasons for their failure. Rao et al. (2012) identified 2 possible factors that affect BPR outcomes negatively: firstly, not taking into account the environment of the process and focusing too heavily on the business process steps and secondly, misalignment between tools for modelling business processes and the automated analysis for identifying the cause of inefficiencies. Sikdar and Payyazhi (2014), on the other hand, named lack of alignment of workflow redesign with organizational change process as one of the major reasons why 70% of BPR initiatives failed. Moreover, Willcocks and Smith (1995) stated that quite often issues with BPR projects arise from them being multidisciplinary and cross-functional. In fact, the broader the scope of implemented BPR projects, the higher the risk of their potential failure, but the concrete level of scope is yet to be specified (Ozcelik, 2010). Among other factors that could contribute to BPR failure there are inadequate choice of team members, inappropriate choice of tools for both developers and users, goals that are not stated in a measurable and attainable way and lack of strong and committed leadership due to the conflict of BPR nature and commonly used leadership style (Sutcliffe, 1999). Successful BPR leaders are typically the ones that adapt their leadership styles according to what is needed in the moment and can find balance between being directive and allowing creative freedom to employees. Abdolvand et al. (2008) proposed to look at critical success and failure factors of BPR initiatives as organizational readiness, which can be assessed in order to predict potential outcome of BPR project. Thus, organizational readiness implies egalitarian leadership, collaborative working environment, commitment of top management, constant management encouragement, and low-to-no resistance to change.

Even though numerous BPR initiatives fail, the ones that succeed help achieve great improvements in business performance. Numerous scientific papers prove the positive effects of BPR. Using panel data analysis, Ozcelik (2010) showed that BPR projects in their completed state significantly increase organization performance in terms of such metrics as labor productivity, return on assets, return on equity, while the implementation process do not affect it in any way. The research implied that such positive effect on firm performance is mostly related to implementation of narrowly focused BPR projects that those with a broader scope. A real-life case study in a form of experience report by Kьng and Hagen (2007) described substantial improvements in 4 different processes of a financial institution after implementing BPR. Overall results have shown that with the use of business process management (BPM) systems, it was possible to reduce cycle time and increase output per employee as well as quality of work products. However, some authors argue that in other industries, such as manufacturing, BPR implementations success is limited, which creates the need for stepping beyond traditional BPR principles and proposing an integrated approach to process automation and optimization based on reengineering (Samaranayake, 2009). Within this integrated approach as proposed by Samaranayake (2009), business processes that concern materials, resources and activities in enterprise resource planning (ERP) can be significantly improved. Thus, companies shall rely on not only BPR itself, but rather enhanced process integration, automation and optimization as a whole.

1.3 Robotic process automation overview

The demand for RPA has indeed increased in recent years, with numerous organizations implementing the technology to minimize costs and redirect employees to more fulfilling tasks. According to Deloitte global CIO survey (2018), process automation and transformation is the primary focus area within digital realm for 69% of companies. Gartner (2018), on the other hand, estimates that 85% of large and very large organizations will apply RPA in some way by the end of 2022. Despite it gaining its momentum in business community, with numerous providers of RPA solutions on the market (Blue Prism, ActiveBatch, RPA Express Automation Anywhere, etc.), RPA is still not thoroughly researched in terms of what it represents for companies and how to successfully implement it. In fact, recent studies of existent literature on the matter showed a high scientific interest in RPA but revealed some significant research gaps in this field. Syed et al. (2020) reviewed 125 articles directly or closely related to RPA from 6 points of view including the concept, potential benefits, capabilities, etc. The review revealed that most of the literature covering RPA topic consists of either position papers or case studies that describe industrial results of RPA implementation (Enriquez et al. 2020) and leaves plenty of room for future research. For instance, Syed et al. (2020) suggested obtaining a deeper understanding and developing a framework of critical success factors of RPA in various business contexts as they may differ from industry to industry. Syed et al. (2020) also proposed developing clear guidelines that might, to a certain degree, guarantee successful RPA deployment in different industries. Moreover, Enriquez et al. (2020) highlighted that it is necessary to further research how RPA impacts organizations as whole. Radke et al. (2020) examined the impact of implementing RPA on different manufacturing companies. Radke et al. (2020) conducted a multiple case study and examined 2 manufacturing companies (one producing electronics, another - pharmaceuticals) that implemented RPA to their master data maintenance activities. The researchers collected information through literature analysis and semi-constructed interviews with representatives of chosen companies. They determined how RPA installation affected the companies. The positive effects include master data quality improvement, decrease in human-made inaccuracies or miscalculations, increase in workforce available for analysis-based tasks and monetary gains in both cases. Based on the findings, Radke et al. (2020) provided practical recommendations for other RPA-interested businesses in a form of a 3-stage-framework: Unfreeze (introduction to RPA and assessing current processes), Change (programming, integrating, cooperating with stakeholders), Refreeze (monitoring progress, correcting issues).

1.4 Robotic process automation effects

Several other studies have been conducted to explore the impact of RPA implementation on productivity, workforce, workload, and attitude towards robotics. Dedrick et al. (2003), for instance, conducted extensive firm-, industry-, and country-level literature research to find out if investments in IT sector affects firm productivity. Investments in both hardware and software are considered IT investment, therefore, investments into RPA technologies (software bots and hardware data storage) could also be viewed as such. The study revealed IT investments, indeed, positively affect productivity on all levels, however some companies experience more prominent productivity gains than others, which might be related to firm capability to deploy IT solutions. Dalebout (2018) conducted semi-structured interviews with project managers and regular employees from four large companies in order to trace any possible alterations in job design provoked by RPA installation. The research showed that RPA deployment produces enrichment and enlargement of job positions involved in the processes that had been automated. Both managers and employees testified that after RPA technology took over repetitive monotonous tasks, new more complex and creative ones had been assigned to the exempted workforce. Employees indicated that a different set of skills was required to complete new tasks as they turned out to be of higher significance and allowed more autonomy than previous ones. Using a similar research strategy, van der Zande (2018) examined the attitude of workers towards RPA technology before and after its implementation. Results of semi-structured interviews indicated that, overall, employees perceived RPA as positive change within the companies and their personal job positions both before and after the project completion. In particular, they noticed significant decrease in the amount of work they were assigned to perform, whereas the tasks became more engaging and thought-provoking.

1.5 RPA-centered reengineering projects and their critical success factors

In general, for reengineering activity to be performed successfully, there must be key enablers within an organization or, as discussed before, an organization must be ready (Abdolvand et al. 2008). Love and Gunasekaran (1997) named several BPR projects enablers: information technology; organizational enabler divided into 2 categories: structural and cultural; human resources and total quality management (TQM). Thus, it could be assumed that BPR projects focused on RPA implementation might as well require a combination of such factors in order to be successful. The IT enabler implies that the organization must possess certain technological capabilities to deploy RPA: data storage, transfer, retrieval and analysis; process mining; robot development, etc (Syed et al. 2020). This implication corresponds with the reimagined BPR roadmap proposed by Deloitte (2017), where at the 3rd phase (Accelerated BPR via Process Robotics) organizational IT potential and capabilities are analyzed to ensure appropriate RPA implementation. Continuing with structural organizational enabler, Love and Gunasekaran (1997) argued that new teams should be established in a firm if the existent ones do not possess skills or experience necessary to deal with process change. Such teams could be self-managing, cross-functional or general-purpose problem-solving ones, however they should be carefully designed to fit into organizational culture. Therefore, RPA-oriented projects might need new team composition where members have adequate experience that fits the requirements and purpose of the projects. Nonetheless, Love and Gunasekaran (1997) emphasized that such work organization might not yield positive results as current situation and other enablers need to be considered as well. The cultural aspect of the organizational enabler, on the other hand, suggests that the organizational culture should be prone to change in order to effectively assist BPR activities. Even more, change management should become “the heart” of BPR (Willcocks & Smith, 1995). Deloitte (2017) takes a similar position and argues that one of the critical success factors for RPA deployment is to overcome the dread of robotics “taking over” an organization and resistance to change. Human resources enabler ensures that workers are still the central focus of the organization since they foster organizational change during BPR activities. RPA is known to change the nature of work for employees because they are offered to switch from repetitive work and focus on more fulfilling duties (Syed et al. 2020). They should be properly trained, skilled and motivated to perform new functions as RPA takes over their previous routinely tasks. Lastly, TQM enabler ensures that all the parties involved into a business process are held accountable and the quality of a final product or service conforms to the accepted standards. Organizations with established TQM are potentially more successful at BPR projects than those without (Love and Gunasekaran, 1997). Especially at BPR projects related to RPA, as TQM can provide clear and strict guidelines for automation.

1.6 Reengineering framework in the context of RPA

Several gaps unsurfaced during the review of the existent literature. Firstly, BPR is rarely studied from the point of view of RPA adoption. Secondly, even if reimagined RPA-centered BPR frameworks exist, they have been created mostly by consultancy companies like Deloitte and are rarely tested by in-house BPR teams. Lastly, such frameworks are generalized and do not provide insight for different business contexts and industries. Considering the gaps identified by conducting a review of prior studies, the research question of this paper is: how has traditional in-house BPR framework changed within the context of RPA implementation in an industry-specific environment. A follow-up question is what the critical success factors of RPA-focused BPR projects are. The main objectives of this research are:

To design the reimagined RPA-centered BPR framework, compare it to existent ones;

To develop industry-specific guidelines for RPA adoption through BPR;

To identify critical success factors for RPA-centered BPR projects.

To achieve such objectives several tasks should be solved: developing a set of questions for interviews; conducting semi-structured interviews; collecting primary and secondary data about BPR activities and RPA implementation in different companies within the same industry; analyzing internal documents and other available data sources.

Prior research analysis revealed several statements regarding the relationship between BPR and RPA that this paper attempts to explore. Table 1 summarizes the identified statements about BPR within the context of RPA and the authors who provided the theoretical basis for such statements. These statements describe the relationship between BPR and RPA and suggest a direction of the research question.

Table 1 - Statements regarding BPR within the context of RPA and their authors

This study is designed to follow the direction that these statements suggest in terms of providing answers to the research questions and achieving research objectives. In order to do that, a qualitative multiple case study is carried out. This research strategy is focused on obtaining knowledge and understanding of the dynamics within specific settings (Eisenhardt, 1989). Case studies are typically carried out to examine episodic events and help answer “how” questions (Chigbu, 2019) by drawing illustrative conclusions (Gustaffson, 2017). Anderson (1993) considered case studies as facilitators for extensive investigation into organizational contexts and the differences between a plan or theory and reality. Noor (2008) specified that the case study is not supposed to examine the whole organization, but rather particular issues or phenomena relevant to research topic. Review of prior research on the topics of BPR and RPA revealed that a great number of them are conducted via case studies (Syed et al., 2020; Enriquez et al., 2020). The case studies helped identify issues when deploying BPR (Fasna & Gunatilake, 2019); test BPR frameworks (Mansar & Reijers, 2005); provided BPR and automation insights from industries like banking (Romao et al., 2019; Kьng & Hagen, 2007) and motorcycle manufacturing (Bevilacqua et al., 2015). Since this study aims at answering the “how” question by designing a demonstrable framework and comparing it to existent ones while adding industry-specific context, the case study research design is deemed as fitting the purpose.

The choice of a multiple case study research strategy can be justified primarily by the choice to show different perspective on the research issue (Creswell, 2013), and to understand differences and similarities between the chosen cases (Stake, 1995). Multiple case studies, when compared to each other, can contribute to the literature with vital remarks about their differences and similarities (Vannoni, 2014;2015). An advantage of a multiple case study, compared to a single one, is that it can provide more plausible results and a more convincing theory, as empirical evidence is supported by multiple instances (Gustaffson, 2017). Other reasons for choosing the multiple case study method are extending the theoretical framework (Yin, 2003) and filling the gaps in existent research on the matter of relationship between BPR and RPA; as well as the need to define industry-specific context for the theoretical framework. Moreover, the suitability of this research design can be justified by the fact that the similar research strategies were used in previous studies (Kowalczyk &Buxmann, 2014; Langstrand & Drotz, 2016; Fasna & Gunatilake, 2019). The selection of cases for this research followed the guidelines proposed by Yin (2003): cases should be selected as fitting for the research topic; they should reflect specific attributes and issues revealed during the prior literature analysis. The companies chosen for the case study are currently leading in iron and steelmaking on both the international and Russian markets. They are among the 9% of Russian companies that have already implemented RPA solutions (Deloitte, 2017) and experienced dramatic performance improvements, have in-house BPR activities and teams and both active and completed BPR projects where RPA tools are still being implemented. Thus, the enterprises possess all the features required to conduct a case study and achieve research objectives.

The industry chosen for the study is iron and steel manufacturing industry. It is often overlooked by researchers since the biggest RPA deployers are typically insurance, banking, utility and telecommunication services (Gartner, 2018), which is also confirmed by prior research analysis that unsurfaced corresponding case studies. However, RPA is widely deployed at steel mills as well, for processing bills of materials (BOM), payrolls and customer orders, transferring data and for any other rule-based and strictly defined processes. An overview of Russian steel and iron market showed that steel production industry is one of the most technologically advanced with the majority of organizations focused on innovating business processes. The survey revealed that 55% of enterprises plan on implementation of RPA solutions, while 9% has already deployed some form of RPA. Nonetheless, the rest 36% has not yet considered RPA implementation (Deloitte, 2017a). Therefore, even though other types of advanced innovative technology are being used, Russian steelworks are still at their early stages of RPA adoption. This study then holds the potential to overview specificities of RPA within steel-making industry and provide experience-based guidelines for future RPA adherents, while RPA implementation continues to increase. Since during BPR activities such tool as benchmarking analysis is commonly used, the findings of this multiple case study can potentially be utilized by companies within iron and steel-making industries for benchmarking analysis for smooth and frictionless RPA deployment.

This multiple case study employed the embedded design as it uses multiple units of analysis or sources of information: individuals and organizational documents (Yin, 2003). Since this research is qualitative, the primary method used to collect the data required was semi-structured interviews. Semi?structured interviews are in?depth interviews where the interviewees respond to open?ended questions that were designed beforehand (Jamshed, 2014). According to Creswell (2013), Hyett et al. (2014), Gaya (2016) in-depth interviews are the most common sources of data used in qualitative case studies. The choice of semi-structured rather than structured interview is justified by the fact that it offers a certain degree of flexibility to find an appropriate approach for each respondent (Noor, 2008), while still following the predetermined topic and the interview guide. This research method is deemed sufficient to achieve research objectives also due to it having been utilized in multiple prior studies on the matter. For instance, Fasna and Gunatilake (2019) used semi-structured interviews to obtain information about issues that BPR adherents faced, while Dalebout (2018) explored the influence of RPA deployment on the involved jobs using the same method. Another method that was employed in this case study is document contents inspection as it is considered appropriate for such research design (Chingbu, 2019). Documentary analysis acts assists in cross-validating information gathered from interviews given that sometimes words and opinions might differ from the actual situation. In addition, documents provide guidelines for conducting more detailed interviews (Noor, 2008).

Considering all the arguments mentioned above, the qualitative multiple case study with an embedded design is deemed sufficient and appropriate to answer the research question and achieve research objectives.

3. Research methods