Wiecher Dalebout – 10679987
MSc. in Business Administration – Digital Business University of Amsterdam – Faculty of Economics and Business
Supervisor prof. dr. P.J. van Baalen Second reader prof. dr. H.P. Borgman
Statement of Originality
This document is written by student Wiecher Dalebout, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is
I would like to thank my thesis supervisor, Professor Peter van Baalen of the University of Amsterdam, for his constructive feedback and support during my thesis. He allowed this paper to be my own work but guided me in the right direction when needed.
Additionally, I would also like to thank my colleagues at KPMG and my supervisors Dirk Kiers and Koen van Raan specifically, for their contribution to this paper with their professional knowledge in the field of RPA and the interesting discussions and useful feedback on this paper.
Finally, I wish to express my gratitude to my (step)parents for their support during my years of study and their encouragement and guidance through the entire process. Without them, this accomplishment would not have been possible, thank you all!
This paper explores the effects of the implementation of Robotic Process Automation (RPA) on the FTE effort and the influence on the involved jobs. Thereby enlightening the field of RPA and building a basis for further research on the topic. The research is based on an interpretivist philosophy and uses an inductive approach, where qualitative data is collected through interviews with managers and employees who are directly involved in the
implementation of RPA. The findings indicate that the implementation of RPA contributes to the enrichment and enlargement of the involved jobs. There were no observations which indicated the existence of a gap in FTE effort. However, the discussed RPA projects have a low maturity and size. The managerial implications of the findings indicate the need for anticipation on the increase in required skills of the employees and the redistribution of tasks after an RPA implementation. Additionally, the findings contribute to the enrichment of the literature on automation by enlightening the benefits and threats of RPA. The limitations of the findings are mainly due to the size of the sample. Finally, the limitations of the paper and recommendations for future research are formulated.
Keywords: Robotic Process Automation, RPA, FTE effort, employment, job design, job enrichment, job enlargement, Human-Machine interaction.
Table of ContentsStatement of Originality ... 2 Acknowledgement ... 3 Abstract ... 4 List of Figures ... 7 List of Tables ... 7 1. Introduction ... 8 2. Literature Review ... 11 2.1 Automation in General ... 11
2.2 Business Process Automation (BPA)... 14
2.3 Artificial Intelligence (AI) & Machine Learning (ML) ... 14
2.4 Robotic Process Automation (RPA) ... 15
2.5 The Job Characteristics Model (JCM) ... 18
2.6 Propositions and Conceptual Framework ... 19
2.6.1 Proposition 1: FTE Effort ... 19
2.6.2 Proposition 2: Job Enrichment ... 20
2.6.3 Proposition 3: Job Enlargement ... 20
2.6.4 Conceptual Framework ... 21
3. Research method ... 23
3.1 Research Philosophy ... 23
3.3 Method and Strategy ... 24
3.4 Time Horizon ... 25
3.5 Respondents Selection ... 25
3.6 Data Collection ... 26
4. Findings ... 26
4.1 Analysis of FTE Effort... 27
4.2 Analysis of Job Enrichment ... 28
4.2.1 Managerial Findings on Job Enrichment ... 28
4.2.2 Employee Findings on Job Enrichment ... 31
4.3 Analysis of Job Enlargement ... 34
4.3.1 Managerial Findings on Job Enlargement ... 34
4.3.2 Employee Findings on Job Enlargement ... 35
4.4 Comparative Analysis of the Findings ... 35
5. Discussion ... 37
5.1 Theoretical Framework ... 37
5.1.1 Proposition 1: FTE Effort ... 37
5.1.2 Proposition 2: Job Enrichment ... 38
5.1.3 Proposition 3: Job Enlargement ... 40
5.1.4 The Research Question ... 41
5.2 Implications... 41
5.2.2 Scientific Implications ... 42 5.3 Limitations ... 43 5.4 Future Research ... 43 6. Conclusion ... 44 References ... 48 Appendices ... 52
Appendix A: Interview guide – Managers (Dutch) ... 53
Appendix B: Interview guide – Employees (Dutch) ... 55
Appendix C: Respondent Interviews... 56
Appendix D: Quotations of the Respondents – Managers ... 57
Appendix E: Quotations of the Respondents – Employees ... 68
List of FiguresFigure 1. Conceptual Framework on the effects of RPA ... 21
Figure 2. Alterations in Job Design: Job Enrichment Versus Job Enlargement ... 22
List of TablesTable 1. The Levels of Interaction Between Human and Computer ... 13
Table 2. The Levels of Intelligence Within Automation ... 17
Table 3. Outcomes Managers ... 36
Human jobs changed radically over the last three centuries due to several industrial revolutions (de Vries, 1994). Each of these automation driven revolutions altered the industries and created fear of human job replacement. Nowadays, the improvement of the digital technologies leads to the next automation revolution; the robotic revolution
(Anagnoste, 2017). Within this robotic revolution, the implementation of technology and robotics into human jobs initiates the same uncertainty about the future of these jobs. Frey and Osborne (2017) find that the future of employment is susceptible to the computerisation, with around 47% of the current employment in the United States at high risk of being
replaced by robotics.
In the recent decades, automation has been complemented by the exponentially increased power of information technology (IT). This increase of technology stirred up the debate on the relationship between IT and productivity (Brynjolfsson & Yang, 1996). In the literature review of Brynjolfsson and Yang (1996), they found few or no evidence of the relationship between IT investments and an increase in productivity, the so-called IT paradox.
Nowadays, within the robotic revolution, there is a development ongoing in the field of business automation where Robotic Process Automation (RPA) is the next step after business process automation. Business process automation is a type of automation which is supported by software to enhance business processes (Scheer, Abolhassan, Jost, & Kirchmer, 2004). Known examples are customer relationship management (CRM), enterprise resource planning (ERP), and supply chain management (SCM). The next step of business automation, RPA, is a relatively new and more advanced form of automation where software robotics are used to automate manual, rules-based processes which rely on structured data and predefined outcomes (Aguirre & Rodriguez, 2017). This main advantages of this form of automation are
the relatively low costs of implementation and maintenance and the fast and easy implementation. With an average implementation time of eight weeks, from high-level process design to the delivered benefits, and cost per robot ranging from €6.000 to €10.000 per year. This cheap and agile implementation, with a rapid increase in productivity, makes RPA a form of IT automation which tends to be less susceptible to the IT paradox.
Besides the simple form of robotic automation, more intelligent robotic solutions exist such as Artificial Intelligence (AI) and Machine Learning (ML). These more intelligent solutions are converging and will enable RPA in the future to apply cognitive tasks as well (Davenport & Kirby, 2016). Considering the ability of AI and ML to interpret, predict, and judge the data within the processes (Agrawal, Gans, & Goldfarb, 2017).
This paper focusses on the effects of the implementation of simple RPA, which is RPA without the application of AI and ML, which is the most used form of RPA since only a few RPA vendors provide more intelligent versions (Le Clair, Cullen, & King, 2017).
Anagnoste (2017) finds that RPA has several challenges and opportunities to deal with. He indicates that one of these challenges is the replacement of employees due to RPA and the uncertainty of this occurrence. This threat occurs in a market which is rapidly growing; in 2016 the RPA market is valued at $250 million, and an estimated market grow towards $2.9 billion in 2021 with over 4 million robots performing administrative, sales and other office tasks (Le Clair et al., 2017).
The literature on automation in general indicates different benefits and threats to the current workforce. With the increasing automation of the human workforce, the importance of the effects of automation become more critical. Where the effects on the number of jobs due to automation are uncertain, however even without the reduction of jobs, these
automations can affect the quality of jobs (Autor, 2015). Since RPA is a rapidly growing form of automation, which is less susceptible to the IT paradox due to the cheap and fast
implementation, the effects of this type of automation on the quantity and type of jobs are more substantial. Therefore, this paper focusses on the effects of the implementation of RPA on the workforce and tries to answer the question:
What are the effects of the implementation of Robotic Process Automation on FTE effort and the impact on the job design of involved jobs within large companies?
This study examines the effects of the implementation of RPA on the alterations of job characteristics of the jobs directly involved in the implementation of RPA. Furthermore, the gap between intended FTE reductions and the realised savings of FTEs is examined to gain insights into the changing job landscape surrounding RPA.
A mono-method study is conducted with qualitative data collection in the form of in-depth interviews with respondents selected from multiple companies. These companies are active within the financial sector, fast-moving consumer goods, and an energy utility
services. Within these sectors, the adoption and maturity of RPA are at a relatively high level. The collected data is summarised to complement the theory on RPA and further emphasise the rigour and depth of the theory (Eisenhardt & Graebner, 2007).
The results of this research will contribute to the understanding of the effects of RPA on FTE effort and gains insights into the alterations of the job design of the involved jobs. These insights contribute to the enrichment of the current literature on RPA and the field of automation in general. Additionally, this research provides recommendations for managers on future RPA implementations and gain insights on the expected effects within their business unit and the implications for their employees.
The remainder of this thesis has the following structure. The second part discusses the relevant literature on the impact of general automation, business process automation, artificial intelligence, and robotic process automation. Additionally, the Job Characteristics Model
(JCM) is discussed and the propositions and theoretical framework are presented. In the third part, the research method and the data collection method are explained. Within the fourth part, the qualitative data is analysed and the main findings are displayed. The fifth section discusses the main findings in relation to the theoretical framework and identifies the
shortcomings of the research, followed by recommendations for future research. Finally, the last section summits the paper and concludes the findings of the research.
2. Literature Review
To explore the field of RPA, the literature on automation is explored from the general perspective of factory robotics since the industrial revolution towards the current landscape of automations. With the focus on business process automation (BPA), artificial intelligence (AI), machine learning (ML), robotic process automation (RPA) and the Job Characteristics Model (JCM). Finally, the chapter is concluded by the three propositions and the conceptual framework on the effects of RPA.
2.1 Automation in General
In the last decades, revolutions in technology have led to fear of job loss and
alterations of current labour (Autor, 2015). This debate on the fear of job loss continuous due to the research of Frey & Osborne (2017), they state that almost half of the jobs in the USA are at high risk of being replaced by computerisation in the coming decades. However, Arntz, Gregory & Zierahn (2017) stress the overestimation of the research of Frey and Osbourne (2017), due to their consideration of the computerisation on occupation-level. Arntz et al. (2017) repeat the same analysis, but use a job-level approach where several tasks within a job are taken into account, and find an automation risk of jobs within the USA of 9%. They also state that the possibility of automation does not imply that automation will indeed take place.
The academic literature thoroughly explorers the influence of automation and the impact on the quality of employment and the number of jobs. Qureshi and Syed (2014) research the influence of robotics in health care and the impact on employment. They state that it is inevitable for organisations to replace human jobs with robots in the service industry and the effects on employees and their motivation are bilateral, both positive and negative. Hence, they emphasise the importance of the improvement of employment and to keep in mind the motivation of employees alongside the implementation of robotics.
Argote and Goodman (1985) address the influence of robotics on individuals and factory companies. They focus on the displacement of employees and alterations for retained staff. In their research, they find likewise opposing effects. Modifications of current jobs, accompanied by in-house training programs, omit some of the potential loss of jobs for current employees. These retained jobs result in positive effects on the employee by learning more skills, performing more significant tasks, and more interaction with colleagues.
However, they find adverse effects for employees who got partly replaced by robots but retained a part of their former job. These employees perform less significant tasks and experience lesser control, and they also experience a decrease in interaction with others.
Autor, Levy and Murnane (2003) further research the overall influence of technological change on the quality and alteration of jobs. They find that computer
technology is a substitute for routine tasks and complementation for non-routine tasks which require cognitive skills. Additionally, they conclude that computerisation results in reduced labour input of routine manual and routine cognitive tasks and an increase in labour input of non-routine cognitive tasks. Moreover, their research explains the effects of tasks shifts and clarifies 60 per cent of demand shift favouring college labour between 1970 and 1998. These findings of Argote & Goodman (1985) and Autor et al. (2003) indicate that the effect of automation is related to the level of automation.
These various insights on the effect of automation indicate that the views within the current literature on automation differ, both in the methodological perspective, the impact on employment and the realisation of the possible implementation of automation. Although it is not an exact science, Parasuraman, Sheridan, & Wickens (2000) outline a model for types and levels automation and to what extent this interacts with humans. Their four types of automation are information acquisition, information analysis, decision and action selection, and action implementation. These levels of automation are based on the model of Sheridan & Verplank (1978), who defined a scale from low (level one) automation, where the human does the entire job, to high (level ten) automation where the computer decides on its own and performs the job itself. Table 1 shows the different levels of the interaction between human and computer.
2.2 Business Process Automation (BPA)
Within a business enterprise, business processes are captured and re-engineered through workflow management (Georgakopoulos, Hornick, & Sheth, 1995). This workflow management can automate processes through Business Process Automation (BPA) (van der Aalst, La Rosa, & Santoro, 2016). Within the field of automation, BPA is a type of
automation which is supported by software to enhance business processes and enables the implementation of best practices at a profitable cost level (Scheer et al., 2004). Examples of BPA solutions are Enterprise Resource Planning (ERP), Supply Chain Management (SCM), and Customer Relationship Management (CRM). This type of automation is a critical link between strategy and execution and is supported by standard software packages to enhance business processes. Therefore, BPA is automation software that mainly focusses on the support for the human within a process (Davenport & Kirby, 2016).
2.3 Artificial Intelligence (AI) & Machine Learning (ML)
A different field of robotics within business process automation is Artificial
Intelligence (AI), where Machine Learning (ML) is part of (Frey & Osborne, 2017). AI is a robotic system which can perform human tasks, such as interpretation of data, prediction, judgment, and action with or without the engagement of a human (Agrawal et al., 2017). ML is a computer process which automatically learns from instances and a set of rules or
classifiers, which are used to generalise to new situations (Kotsiantis, Pierrakeas, & Pintelas, 2003).
In their research, Agrawal et. Al (2017) state that the critical challenges for the future are the adjustment of the training of employees, to estimate the speed and direction of the adoption of AI to determine the shifting of the workforce, and the development of most efficient teams with a combination of humans and AI. Additionally, Frey & Osborne (2017)
underline the challenge for human employees to adjust to automation with the addition of AI, due to the learning capacity of the AI. Likewise, Chelliah (2017) also states the importance for human resources departments to prepare for the impact of AI on the current workforce regarding the type of work and the role of employees in the optimal mix between robotics and humans. Liebowitz (1989) emphasises the importance of the alignment of AI with the goals and objectives of the firm and the available resources. Furthermore, he states that the overall effect of AI will be positive, but the impact on society needs to be measured and anticipated. He underlines that the focus needs to be on the impact of AI on the creation, alteration and replacement of jobs. Arntz et al. (2017) debilitate the impact of automation on human jobs. They find that workers increasingly focus on diversification and therefore complement the implementation of technologies. Although their research still indicates that 10% of the jobs are highly exposed to the implementation of robotics, they emphasise that the overall job losses depend on the relative numbers of job creation and job destruction.
Davenport and Kirby (2016) discuss these different types of automation and indicate the differences from simple automation towards a more complex cognitive form of
automation. In their paper, they also state that the fields of robotics and AI are converging, which indicates that simple process automation becomes more intelligent. Some RPA
vendors, such as WorkFusion and Pegasystems, are combining RPA and AI in their products (Brynjolfsson & Mcafee, 2017; Le Clair et al., 2017). However, this paper focusses on the simple form of RPA without the addition of AI or ML, since this is the most prominent part of the current RPA market (Le Clair et al., 2017).
2.4 Robotic Process Automation (RPA)
A new and revolutionary type of business process automation is RPA, which is considered an uncomplicated form of AI (Anagnoste, 2017). This form of automation is
considered revolutionary due to the ease of use, the low price and the fast implementation. Where RPA is being defined as a software-based solution for the automation of rules-based processes that consists of repetitive manual tasks with structured data and pre-determined outcomes (Aguirre & Rodriguez, 2017). Moreover, RPA is a software application used to interpret and capture existing applications with the goal of communication across multiple IT platforms, data manipulation and transaction processing (Suri, Elia, & Van Hillegersberg, 2017). In this paper, RPA is defined as the latter definition. Additionally, although the possibility of the application of AI on RPA exists, it is not included in this paper, since the integration of AI on RPA is uncommon (Le Clair et al., 2017). Due to the pre-determined outcomes and the installation of the bot by humans, RPA can be scaled in level seven of interaction within the model of Parasuraman et al. (2000), which is displayed in Table 1.
There are two main differences between the implementation of RPA and classic business automation (Lacity & Willcocks, 2015a). Firstly, programming of RPA can be learned with a few weeks of training; hence no extensive programming experience is needed. Which results in a cheap form of automation with a quick way to achieve a high return on investment (van der Aalst, Bichler, & Heinzl, 2018). Secondly, RPA automates a process with an "outside-in" approach; therefore it controls the computer on the user interface level, which does not disturb the underlying computer systems (van der Aalst et al., 2018). These main differences provide a significant benefit over traditional business automation, which is done according to an “inside-out” approach. Besides the quick achievement of high return on investment, the primary goals of the implementation of RPA are; cost reduction, quality increase and faster processes (Anagnoste, 2017). Additionally, the implementation time of RPA has an average of eight weeks from the high-level process design until the delivered benefits. Due to the fast implementation, the low cost, and the rapid increase in productivity, RPA tends to be less susceptible to the IT paradox.
Moreover, in Table 2 the differences in levels of intelligence between the several forms of automation are explained to create an overview of the current tools within the field of automation.
The Levels of Intelligence Within Automation (Davenport & Kirby, 2016)
These benefits of RPA over classic business automation lead to a rapidly growing RPA market. With a total market value of $250 million in 2016, and an estimated growth towards a market value of $2.9 billion in 2021 (Le Clair et al., 2017). Furthermore, it is estimated that around four million robots will perform rules-based digital tasks within the automation market in 2021.
The current literature on automation in general provides different insights into the benefits and threats of automation on employment. Since a growth of the automation market is expected, the effects of automation on the current workforce become increasingly
important. Especially, the prediction of Frey and Osbourne (2017) that 47 per cent of jobs are susceptible to the computerisation and therefore emphasise the importance to explore the effects of automation on the current workforce.
Within the field of automation, RPA is a rapidly growing software tool within the growing market of automation. Also, due to the novelty of RPA, there is uncertainty about the effects of RPA on involved jobs and the affected processes. The literature indicates that the implementation of robotics in the workforce can influence the employees both positively and negatively. Although plenty of research is conducted on AI in general, RPA is a more straightforward version and therefore different form of automation. Since this is a different type of automation, further research is needed to determine the effects of RPA on the quality and the quantity of labour (Lacity & Willcocks, 2015b). Implementations of RPA within large companies are researched to provide an answer to the following research question:
What are the effects of the implementation of Robotic Process Automation on FTE effort and the impact on the job design of involved jobs within large companies?
This research question is formulated to explore the phenomenon of RPA and therefore contribute to the current literature on the field of automation and the influence of robotic automation on employees. Additionally, this exploratory study gains managerial insights into the overall effects of robotic automation on employees in practice, which is supportive of managerial decisions whether or not to apply robotic automation within a business process.
2.5 The Job Characteristics Model (JCM)
As a basis for the conceptual framework of this paper on the impact of the job design of the involved jobs, the JCM of Hackman and Oldham (1976) is used. This model explains the relation between intrinsically motivating characteristics and personal work outcomes. They determine five key job characteristics: skill variety, task identity, task significance, autonomy and feedback, which all positively influence the work motivation, performance, satisfaction, and turnover of the employees.
Despite the age of the Job Characteristics Theory and the JCM on job design, it is widely used and accepted in recent literature. Miner (2015) analyses the model and concludes that the model has undergone several expansions to increase the precision of the predictions of the model. Furthermore, Miner (2015) describes the model as highly important and useful, and above average on validity. This support for the model is emphasised by the number of recent papers which are based on the JCM of Hackman and Oldham.
2.6 Propositions and Conceptual Framework
The following three propositions and the conceptual framework are designed to explore the phenomenon of RPA and are used as a guide to answer the research question. The first proposition focusses on the effect of an RPA implementation of the FTE effort. The second and third propositions focus on the effects of an RPA implementation of the job design of the involved jobs, respectively concerning job enrichment and job enlargement. These propositions and the framework form the basis of the research design and the interview protocol.
2.6.1 Proposition 1: FTE Effort
The disagreement between the research of Frey and Osbourne (2017) and the paper of Artnz et al. (2017) indicates a misalignment between the savings in FTEs as a result of robotics. While Frey and Osbourne (2017) state that 47% of the jobs in the USA are at risk of being replaced by robotics, while Artnz et al. (2017) estimate this risk on 9%.
This overestimation of savings in terms of FTEs is also expected within the field of RPA, leading to the following proposition.
Proposition 1: The realised FTE reduction due to the implementation of RPA is lower
2.6.2 Proposition 2: Job Enrichment
The RPA bots are designed to perform repetitive rules-based tasks with the use of structured data (Aguirre & Rodriguez, 2017). These bots are predominantly designed to take over the routine tasks; hence it is expected that the employees will substitute these routine tasks for non-routine tasks (Autor, 2015). Moreover, it is expected that these non-routine tasks require an increase in skills, a more extensive task identification, an increased
significance, and more autonomy. However, it is expected that the feedback element of the JCM will remain equal since the RPA implementation does not directly alter the link between the manager and employee regarding feedback. Concluding, the expected effects of RPA on the involved job design concerning skillset, task identification, task significance and
autonomy, results in the following proposition.
Proposition 2: The implementation of RPA leads to job enrichment of the involved jobs.
2.6.3 Proposition 3: Job Enlargement
After the implementation of RPA, a set of routine tasks is performed by a bot instead of a human employee. However, due to the implementation of RPA, it is expected that new tasks occur, for example, the supervision and error handling of the bot. Furthermore, the decrease in routine tasks provides the possibility to perform more non-routine tasks. Besides the proposed enrichment effects, this dislocation of tasks is expected to influence the
involved jobs in term of width as well. Where the enlargement of the job is determined as the task variety and the number of tasks performed by the employee (Chung & Ross, 1977). Therefore, the following proposition is designed.
2.6.4 Conceptual Framework
To gain insights into the proposed effects of the three propositions, the framework in Figure 1 is constructed. The first part of the framework illustrates the transformation of the human workforce, with the intended FTE reduction, towards the new situation with the combination of human and digital workers and the realised FTE reduction. Proposition 1 is graphically displayed as ‘the gap’ between the intended and realised savings in terms of FTEs. The second part of the framework displays the elements of the job design of the remaining workforce. Within the job design of the involved jobs, an alteration is expected, this is both possible due to job enrichment (P2) and job enlargement (P3) or one of both. The Job Enrichment effects are explored based on the JCM (Hackman & Oldham, 1976). The enlargement of the involved jobs is explored according to the alterations in number and variety of tasks performed by employees due to the implementation of RPA (Chung & Ross, 1977).
Within this framework, the two different main job dimensions reflect job enrichment and job enlargement, where job enrichment is about the depth (vertical axis) of a job, hence the quality. While, job enlargement is about the breadth (horizontal axis) of a job, concerning an increase in duties and workload, hence the quantity (Chung & Ross, 1977). The possible alterations from the current job situation are illustrated in Figure 2. This figure displays the current situation of the job design within the middle circle and the potential alterations from this middle point.
3. Research method
The purpose of this research is to explore the phenomenon of RPA and the effect of the implementation of RPA on FTE reduction and the potential alterations of the job design of jobs directly involved by RPA. The research on FTE reduction focusses on the gap
between intended FTE reduction before the implementation and the realised savings on FTEs after the implementation. The alterations of job characteristics are examined according to the conceptual framework, which is distracted from the JCM (Hackman & Oldham, 1976) in combination with the theory on job enlargement (Chung & Ross, 1977). This conceptual framework is used to explore the impact of RPA implementation on the remaining employees their job characteristics. Which leads to the following research question:
What are the effects of the implementation of Robotic Process Automation on FTE effort and the impact on the job design of involved jobs within large companies?
This chapter describes and justifies the research design and techniques applied in this thesis to answer this research question. The focus lies on the chosen research philosophy, approach, strategy, method, time horizon and the data collection method.
3.1 Research Philosophy
To explain the choices of the researcher regarding the selection of the proper
approach and strategy for the research, it is essential to gain insight into the philosophy of the researcher (Saunders, Lewis, & Thornhill, 2012). This philosophy describes the view of the researcher on the world.
Through this study, it is assumed that people are social actors who base their behaviour on their interpretation of the world around them; thus the research philosophy of
this paper is interpretivism. In this respect, the researcher is in the social world of the researched subject to gain explore the phenomenon from the right perspective.
Due to the novelty of the topic of this research question, the chosen philosophy is backed by an inductive approach, where the observations are used to explore, describe, and explain the phenomenon (Saunders et al., 2012). This approach is chosen to explore the effects of RPA since it is a relatively new phenomenon within business process automation. This approach aims to explore the effects of RPA on the FTE effort and the job design of the involved jobs. Where the findings can be seen as a foundation for new propositions,
hypotheses, and theories to build from in future research.
3.3 Method and Strategy
Since the goal of this research is to emerge new theory from its findings, a qualitative method is chosen. This method is useful when a new phenomenon is researched, with limited or vague literature available (Blumberg, Cooper, & Schindler, 2008). Besides, this method is often used to explore the field of a concept and further develop the theory of a phenomenon.
This qualitative data is collected using a mono-method study over multiple
respondents. These respondents are selected from two different types of employees for the interviews. The first type of respondents are the process owners, on a managerial level, where the focus of the interviews lies on the alterations of job characteristics of their employees and the potential gap between intended and realised FTE reduction due to the implementation of RPA. The second type of respondents are the employees who are directly influenced by the implementation of RPA. These interviews focus on the effect of the RPA implementation on the possible alterations of their job characteristics. These respondents are selected within different companies, which result in an exploratory study conducted over multiple scenarios.
This approach is chosen to investigate a phenomenon within its real-life context to answer the ‘why’ and ‘how’ questions (Yin, 2003).
3.4 Time Horizon
The chosen time horizon for this research is predetermined and is set between January 2018 and May 2018. This time horizon is convenient chosen, due to the limited time
available for this research. Within this timeframe, the data is collected in a cross-sectional nature, at a certain point after the implementation of RPA to research the a posteriori effects of RPA. This nature is chosen since it is expected that the respondents have a clear view after the implementation of the effects that emerged from it.
3.5 Respondents Selection
The respondents are chosen from predetermined companies according to a set of requirements to provide a clear answer to the research question. One of these requirements for the companies is that the implementation phase is completed and RPA is up and running. This completion of the implementation is required since the research focusses on the a
posteriori effects of the implementation of RPA. Additionally, the process needs to be owned
by a manager and previously or currently executed by an employee to explore the effects of the RPA implementation on the job design of the employee. Finally, the company size is set at large, with more than 250 employees to collect data which is more comparable and generalizable. Another reason for selecting respondents from large companies is due to convenience since all of the customers of KPMG are large companies. Concluding, the respondents are selected using a mix of purposeful and confidential sampling since the respondents are selected according to the criteria of their company size and the RPA
implementation maturity of the company, and according to the availability and willingness of the companies to cooperate.
3.6 Data Collection
Semi-structured interviews are conducted to gain insights into the possible effects of RPA implementation on FTE effort and the job design of involved jobs. The semi-structured interviews result in rich and unanticipated data, which is the better of two worlds (Saunders et al., 2012). Two interview guides are designed according to the three propositions and the conceptual framework. These guides are used to gather the correct information to explore the field of RPA and answer the research question. The guides are displayed in Appendix A, for the managerial interviews, and in Appendix B for the employee level respondents. Moreover, the interviews are conducted at the offices of the respondents, in a confined space, to make the respondents feel comfortable so that they can speak freely.
In this chapter, the observations of the respondent and the analysis method are described. Followed by the analysis of the findings per proposition of the conceptual framework and respondent type. In total ten qualitative interviews are conducted, five at managerial level; respondents M1-M5, and five at work floor employee level; respondents E1-E5. These interviews were conducted at five large firms; two banks, one company within the fast moving consumer goods sector and two energy utility companies. These firms are all Dutch-based companies and categorised as large, with over 250 employees. Additionally, RPA has been implemented and is still active at all of the processes which were discussed during the interviews. An overview of the respondents and the corresponding sectors of their company is displayed in Appendix C.
The collected data is analysed by transcribing and coding the interviews. The transcribing of the interviews is done manually to guarantee the literal transcription of the respondents their answers. Afterwards, the collected data is thoroughly analysed by coding
the transcriptions. From this analysis, the findings are described in the following sections per proposition and separately per group of respondents. The overview of the quotations are displayed per type of respondent in the appendices; Appendix D for the quotations of the managers and Appendix E for the quotations of the employees.
4.1 Analysis of FTE Effort
The potential gap on FTE effort is questioned during the managerial interviews. Throughout these interviews, all of the respondents indicated that there was no gap between the intended and realised FTE effort of the discussed RPA project. Only respondent M1 mentioned a small indication of a difference between the business cases of the project. He indicated a gap between the first business case of a saving of 0.6 FTE, against an adjusted business case of 0.2 FTE. This adjustment was made before the implementation of RPA; therefore it does not indicate an FTE gap. However, the misalignment between the two business cases does show some indication of a possibility between the intended FTE savings on the forehand, and the realised FTE savings after the implementation of RPA.
Furthermore, respondent M2 and respondent M3 both indicated that the realisation of FTE reduction is difficult, where respondent M3 states that: “(…) not all of the projects are
without a gap, but this [successful project, without a gap] is a nice example”. As an
explanation, both of the respondents indicate the difficulty of realising short-term FTE
reductions on business cases below one FTE. They indicate that other work is allocated to the FTE where time is made available due to RPA. Furthermore, respondent M3 explains that the gap in the mentioned process does not exist, he explains: “I had, of course, given the figures
and arranged for the implementation, so in the end, you have to set a good example.” Finally,
respondent M5 indicates that there is no difference between the intended and realised savings in terms of FTE, or no data is available. He indicates that due to the novelty of RPA, the first
projects have the purpose of training the staff in the use of RPA and explore the outcomes of an RPA implementation. In the future, his company will scale up the projects towards projects of the size of reductions of 60 FTE instead of the current projects with a size of one or two FTEs. Concluding, the data collected show no support on the first proposition on FTE effort. Only some indications that the eventual realisation of FTE reduction can be
challenging are pointed out.
4.2 Analysis of Job Enrichment
Both managers and employees provided information on the effects of RPA implementation on the job design of the involved jobs. The gathered information on
Proposition 2 about job enrichment is discussed according to the five dimensions: skills, task identity, task significance, autonomy and feedback. The managers provide insight into the effect of RPA on the involved jobs of their employees, while the work floor employees share their experience on the possible alterations of their job.
4.2.1 Managerial Findings on Job Enrichment
All of the managers noticed some form of job enrichment on the job design of their employees, due to the implementation of RPA. This enrichment was indicated when
discussing job enrichment in general. However, the five dimensions of job enrichment were further discussed separately, and the findings are displayed per dimension.
The observations of the managers on the required skills of their employees due to the implementation of RPA were unambiguous. All of them indicated that the implementation of RPA led to higher required skills to perform their adjusted jobs. Respondent M1 states: “It
makes your job more challenging because you no longer have to do your standard bulk work", referring to the jobs of his employees. Where respondent M2 describes the alterations
indicates an increase in the required skills for the function. Furthermore, respondent M4 anticipated on the RPA implementation by hiring higher educated staff; Secondary Vocational Education (SVE) level five instead of SVE level four. He states that the main difference between those scales is the capability to think along in the process instead of only performing the job which is asked of the employee. Additionally, respondent M5 indicates that the possibility exists that both SVE level four and five will dissolve due to the
implementation of robotics, which results in a financial function which requires SVE level six staff. However, he further states that these expectations have been around in the previous revolutionary automation waves, so it is uncertain if this prediction will come true.
The managerial view on alterations of the task identity sub-dimension indicates no alterations due to the implementation of RPA. Only respondent M2, M4 and M5 state an increase in task identity. However, M5 explains that this increase is partly due to a lean transformation within the organisation where employees get to be part of the entire end-to-end process instead of just a small part of the product. Respondent M1 states that the RPA bot takes over a time consuming, but a small part of the end product and therefore the task
identity of the job of the employees do not alter. The last respondent, M4, indicates that he does not observe alterations within the task identity of the job design of his employees.
Furthermore, four out of five respondents noticed an increase in the task significance of the employees their job design. Respondent M2, M3, and M4 all observe a task shift towards work which the department wanted to do before, and now the possibility arises for his staff to perform these more complicated and more critical tasks. On the other hand, this development cannot be entirely dedicated to RPA, as stated by respondent M2, who indicates that part of this development is conjointly due to other projects. Respondent M5 does not state an increase in the significance of the tasks, although he does mention that the tasks add
more value to the company, he indicates that all the tasks within an organisation are of equal significance, from top management towards the repetitive administrative tasks.
Within the data on the job enrichment dimension of autonomy, all respondents report an increase in the autonomy of their employees due to RPA. Respondents M3 observes an increase in autonomy due to the replacement of routine tasks towards tasks which the
employee can decide to carry out on their initiative. Respondent M4 emphasises this increase in autonomy; he indicates that his employees require less control within the new setting. Respondent M5 states that the current repetitive tasks that can be replaced by RPA are tasks with little autonomy. He observes that the shift towards more challenging and value-added tasks will increase the autonomy of the employees.
Possible alteration of the job design in the final dimension of job enrichment,
feedback, is not experienced by the respondents. Four out of five respondents indicate no
alteration in their feedback towards the employees due to the implementation of RPA. Where respondent M3 states: "RPA is just a tool, so I do not see why I should alter my feedback
towards my employees, so in my perspective, the feedback should not change due to RPA".
Respondent M4 does observe alterations in the feedback on employees, but he mentions that all the dimensions alter and lead to job enrichment, without a specific explanation of the feedback dimension. Respondent M5 indicates that no alteration in feedback occurs due to the implementation of RPA, he does emphasise that the applied lean method changes the feedback system since employees provide feedback to each other based on statistical data. Concluding, none of the respondents observes an alteration in the amount of feedback that is received by their employees due to the implementation of RPA.
With all the information on the five sub-dimensions of the JCM on job enrichment, it is concluded that all of the interviewed managers observe an overall enrichment of the
involved jobs due to the implementation of RPA, both in perspective on the sub-dimensions as in the general observations on job enrichment.
4.2.2 Employee Findings on Job Enrichment
The observations of the employees on job enrichment in general and the five dimensions separately provide insights into the view of the employees on the impact of the implementation of RPA. When focussing on the alterations of Job Enrichment in general, three out of five respondents experience an enrichment of their job due to the implementation of RPA. Furthermore, the employees experience less enrichment within the separate
dimensions compared to the managerial view on the effect of RPA on their job design. The employees their view on the required skills due to the implementation of RPA differ. Three out of five experience no alteration in the required skillset within the new setting, while the other two respondents describe contradictory results since they do notice an apparent increase in the required skills. Respondent E1 sees no alterations in the required skills for the job; he states that the tasks are different, but not more complicated than the previous tasks. Respondent E2 first indicates that he sees no increase in the needed skills for the job, however, at a later stage during the interview, he indicates that his job complexity increased due to the implementation. He is assigned more complex tasks which require higher skills. Respondent E3 indicates a similar contradictory view on the skillset; on the one hand, the respondent states that he already had the knowledge of all the tasks within his department. However, he also states that he gained new skills in Microsoft Excel that he could not
perform before. Respondent E4 observes that more skills are needed after the implementation of RPA, he states: “Now you receive the rapportage made by the robot, and you need to be
able to analyse it and understand the possible mistakes made by the bot". He further states
due to errors in the delivery by customers. These errors need to be corrected, and the cost of this error handling is recharged to the customers. This new role requires a higher skillset because the entire process must be understood and the correct billing for the robot hours is needed. Overall the employees see some alterations, but it seems that it is difficult to
precisely point out the alterations in their required skills due to the implementation of RPA. The respondents their view on the task identity dimension furthermore indicates different experiences. One employee, respondent E1, indicated that he did not experience any alteration in the depth of his job. However, he did see some minor improvements on task identity, because of his new role of the implementation of other RPA bots within their department. Therefore he is more intimately involved with the entire product of the
department. Respondent E3 also experiences this increase in task identity; he states that his contribution to the entire process has increased since the robot is installed and he takes part in multiple parts of the end product. Respondent E4 explains that the robot takes over a part of his job, and therefore he observes a decrease in the task identity. While respondent E2 and E5 do not experience an alteration in their participation in the entire process.
The observations of the employees on task significance indicate an increase of this sub-dimension due to RPA. Four out of five employees experience an increase in the significance of their tasks. Only respondent E1 does not experience an alteration in the significance of his tasks since the setup of RPA. Contrary, respondent E2 states that he performs more complicated and more important tasks now, which is also experienced by respondent E3, who has a feeling that his job became more important. Respondent E4 links the significance of the new jobs to the RPA bot since the robot is now operating
autonomously, it becomes more significant to monitor the errors of the bot and deal with those mistakes adequately. Respondent E5 indicates that the increase of significance of his job is due to his role in the implementation process of the robots.
The observations of the employees on the autonomy dimension indicate a minor influence of the RPA setup. One of the employees, respondent E3, indicates an increase in the level of autonomy in his job after the implementation of RPA. He explains the regular daily task norm they have to achieve, due to the implementation of RPA he can spend more time on day-to-day tasks that come by, where he can operate more autonomously. Three other respondents, E1, E2 and E5, observe no increase, nor a decrease, in the level of autonomy of their job due to the RPA bot. At last, respondent E4 observes an adverse effect of autonomy due to the standardisation of his current job. He indicates that a new monitor role emerges, where standard procedures are followed in case of an error of the bot.
Within the feedback sub-dimension, the employees experience no alteration due to the RPA bot. They indicate that they do still receive feedback, but the same amount of feedback on the new tasks they are performing. Respondent E3 does stress that the feedback became more important: "So feedback, especially lately, has become more and more important within
the organisation and also within our department itself. It was difficult to give at first (…), but I think that we have now broken that barrier at this moment”. However, this does not indicate
an alteration in the amount of feedback received.
The employees view on the general enrichment of their jobs is positive for three respondents, while respondent E1 and E5 do not experience any alterations in the depth of their jobs. Respondents E2 and E3 indicate a more in-depth job due to the available time created by the RPA bot and the new work that emerges from this available time. Respondent E4 indicates that his job is enriched since he got the opportunity to support the RPA project by selecting and explaining other processes, however, he also indicates that the roles of employees who only need to monitor the bot did not experience an enriched job. Concluding, the effects on the enrichment of the employees their jobs differ per case, the time that comes available after the implementation of RPA can result in a job enrichment, but this is not
guaranteed. The new tasks that emerge from the implementation itself can further result in a more in-depth job.
4.3 Analysis of Job Enlargement
Both managers and employees shared their observations of the effect of the
implementation of RPA on the enlargement of the involved jobs. The gathered observations on Proposition 3 are discussed according to the alteration in variation and amount of tasks of the involved jobs. The managers provide insight into the involved jobs within their
department, where the work floor employees share their experience on their job.
4.3.1 Managerial Findings on Job Enlargement
Four out of five of the managers indicated that they observe an enlargement of the jobs within their department. Although respondent M1 states that this enlargement is not entirely due to the implementation of RPA. He explains that RPA is solely the tool to reach the goal of reducing FTE and changing the department, thereby RPA only accelerates the effect. The other respondents do notice a broader function that can be devoted to the RPA implementation.
Both respondent M3 and M4 indicate that the implementation of RPA results in a more extensive variety of tasks and an increase in the number of tasks performed, and therefore a broader and enlarged job. Respondent M3 explains: "There is a shift in the in the
tasks you need to carry out, this indicates more complex tasks and also a broadening of the tasks you have to perform (…) yes, also in the variation of the tasks”. Respondent M5 does
not notice an alteration in the breadth of the involved jobs, he states: “No, I do not see an
enlargement of the tasks of the employees, we focus on narrowing down tasks, so when tasks are taken over by the robot, we fill in this available time by other tasks if needed, but we do not add extra tasks”.
Combining the results on both sub-dimensions of job enlargement, the findings indicate that most of the managers notice an enlargement of the job design due to the RPA implementation. However, this is dependent on the strategy of a company concerning the divisions of tasks, since no enlargement is experienced in the scenarios where no extra tasks are assigned after the implementation.
4.3.2 Employee Findings on Job Enlargement
The employees experience more flared effects of the RPA implementation on the enlargement of their jobs. Respondent E1 acknowledges the addition of some new tasks, but no alteration in the variety of his work, and therefore he does not experience an enlargement of his job. Where respondent E2 and E3 both experience an enlargement in their jobs due to the addition of the RPA bot, with a shift towards more tasks and more variation in their job. Respondent E5 does not notice any alterations regarding the breadth of his job; he states: "I
do not see any more variation, I do the same tasks as I did before, so the introduction of the robots did not provide me with new tasks”. Moreover, respondent E4 experiences a
narrowing of his job, due to the standardisation of his new role, which results in less variety and amount of tasks. He states: " I think that my function became even more narrow and
standardised, which makes it less interesting for people who want to learn something new".
These contrary experiences emphasise the differences per respondents concerning job enlargement, where in some cases employees are assigned to new tasks, and in other cases, employees do not notice any alterations or even a more narrow function within the new situation.
4.4 Comparative Analysis of the Findings
The results of the previous sections combined indicate that the implementation of RPA does affect the current job design of the involved jobs in various ways. Where all of the
managerial level respondents indicate both enrichment and enlargement of jobs, though, not within all of the sub-dimensions of these core job dimensions. These findings indicate the importance of a reconsideration of job design when implementing RPA and the consequences for the current and future workforce.
On the other hand, no support is found for the FTE effort gap. This lack of support for the FTE 'gap' indicates that the correct estimation can be made on the forehand with the implementation of RPA. However, further research on the potential gap is needed to exclude the possibility of the self-serving bias of the respondents, since the gap can be interpreted as a mistake by the manager. This self-serving bias could occur due to the involvement of the manager in both phases of the estimation of the FTE effort.
The overall results are graphically displayed in the following tables: Table 3 on managerial outcomes and Table 4 on employees their outcomes.
Note. A * indicates that the sub-dimension is not discussed separately from the main dimension.
In this chapter, the overall findings of the paper are discussed according to the three propositions of the theoretical framework. The combined findings of these propositions are used to answer the research question. Afterwards, the implications that ensue from these findings, in both managerial as scientific nature, are debated. Thereupon, the limitations of this research conclude the discussion.
5.1 Theoretical Framework
5.1.1 Proposition 1: FTE Effort
The observations about FTE effort indicate that no gap exists within the projects mentioned by the respondents. However, as indicated by respondent M5, the main reason for the indicated lack of support on the FTE gap can be due to the level of maturity of the
discussed projects. As respondent M5 indicates: "This potential difference is not yet visible
have not yet started on the bigger projects of around 60 FTE. In those projects, it is more relevant to measure the realised FTE savings". This statement is in line with the information
provided by other respondents since only small projects, of less than two FTE, were discussed. This lack of volume in the discussed RPA projects can indeed contribute to the absence of the potential gap on FTE effort. Furthermore, since the involved managers are responsible for the realisation of the initial business case, the observations can suffer from a self-serving bias. Since the respondents were all responsible for the project and worked together with KPMG on the business case. This potential bias is accentuated by respondent M3, as he indicates that he was both responsible for the initial business case and the
However, within this research, the observations of the respondents do not imply any support for the first proposition on FTE effort. Therefore proposition 1; the realised FTE
reduction due to the implementation of RPA is lower than the previsioned FTE reduction, is
perceived not to be true.
5.1.2 Proposition 2: Job Enrichment
The findings on the effect of RPA on job enrichment indicate overall positive observations by the respondents on the enrichment of the involved jobs. However, the observations on the sub-dimensions of the enrichment of a job, namely: required skills, task identity, task significance, autonomy, and feedback, are not all convincingly increased due to RPA. Furthermore, the overall observations of the employees on job enrichment indicate more disparate effects compared to the findings of the managers.
Regarding the first dimension of job enrichment; the required skills, the results of the managers provide an unambiguous and clear vision, all of which report an increase in the required skills of their employees as a result of the implementation of RPA. While the
employees experienced a more divided effect on their required skills within the new setting, but all of the employees who noticed an overall enrichment of their job also experienced an increase in required skills.
The observations on the second dimension of job enrichment; task identity, indicates conflictual outcomes, with only three out of ten respondents indicating an increase in the task identity of the involved job due to the implementation of RPA. Therefore, the involvement in the end-to-end process remained unchanged in most of the observed scenarios, since the RPA implementation only took over a small part of the process.
The observations on the third dimension of job enrichment; task significance,
demonstrates an increase due to the implementation of RPA by eight out of ten respondents. This increase is mainly caused by the increased complexity of the new tasks which are assigned to the employees. Most of the managers indicate that these new and more complex tasks are indeed more significant compared to the repetitive tasks that have been taken over by RPA. The employees also experience this increase of importance within their new roles, since four out of five employees indicate an increase in significance.
The observations on the fourth dimension of job enrichment; autonomy, indicate an increase within the new setting according to the managers, they all indicate that the autonomy of their employees increased. This increase is again linked to the repetitiveness of their previous tasks since these repetitive tasks allow little autonomy. The shift towards more complex tasks results in more independence within the involved jobs. This increase in autonomy is not experienced by the employees. Three out of five employees do not experience an increase in autonomy, while one employee observes more autonomy in his work. The fifth respondent experiences a decrease in the amount of autonomy. This contrary results between the observations of the managers and employees indicate an issue that needs
to be addressed since the alignment on perceived autonomy seems to differ from the given autonomy.
The fifth dimension of job enrichment; feedback, does not alter according to all of the respondents. The respondents indicate that the implementation of RPA does not alter the manner of feedback from the manager to the employee. Furthermore, the assignment of new tasks neither alters the feedback on the work of the employees.
With all of the sub-dimensions of job enrichment combined and the responses on job enrichment as a whole, there is a strong indication that managers and employees both
experience an enrichment of the involved jobs; therefore the observations indicate support for the proposition on job enrichment. Hence, Proposition 2; the implementation of RPA leads to
job enrichment of the involved jobs, is considered to be true.
5.1.3 Proposition 3: Job Enlargement
The respondents their insights on the effect of RPA on job enlargement indicate that the shift from repetitive tasks towards more complex tasks increases the number of tasks and the variety of tasks. Most of the managers experience this broadening of the involved jobs; they indicate that the available time is assigned to other jobs which increase the number of tasks and variety. However, when no new tasks were assigned to the employees, no alteration or even a more narrow job was experienced by the employees.
Therefore, the enlargement of jobs which are involved by RPA is linked to the new tasks assigned, where cases with no new tasks assigned result in a narrowing of the jobs, while jobs with the assignment of new tasks are observed as broader jobs. Hence, some observations appear to support Proposition 3. However, this support of the proposition is conditional on the assignment of new tasks to the employees after RPA is implemented. Hence, Proposition 3; the implementation of RPA leads to job enlargement of the involved
jobs, is considered to be true. However, this is conditional on the assignment of new tasks to
the employees after the RPA implementation.
5.1.4 The Research Question
With the discussion of the three propositions combined, an overview is created on the general research question of this thesis: What are the effects of the implementation of Robotic
Process Automation on FTE effort and the impact on the job design of involved jobs within large companies? Due to the findings that indicate no observations of a gap in FTE effort
within the RPA implementations, Proposition 1 considered not to be true. However, the observations on Proposition 2 and Proposition 3 both indicate support for the impact of RPA on the job design of the involved jobs, both in terms of job enrichment and job enlargement.
These findings indicate some support for the research question, where it is considered to be true that; the implementation of RPA impacts the job design of the involved jobs within large companies, both concerning job enrichment and job enlargement.
In this part, the implications of the research in both managerial as scientific context are discussed. First, the professional implications are discussed, focussed on the managers who are involved in RPA projects or will be involved in the future. The scientific
implications of this research succeed this discussion.
5.2.1 Managerial Implications
The findings that RPA affects the jobs design of the jobs involved by the automated process, both in job enrichments as job enlargement, underlines the importance of the
manager his role in this implementation. These alterations of the job design are relevant, due to the positive effect on employees in terms of work motivation, performance, satisfaction,
and turnover (Hackman & Oldham, 1976). To reach the full potential of these positive effects, both the expected enrichment and enlargement of the current jobs need to be addressed when RPA is applied within a business department. This urgency to correctly address these alterations is indicated by the different observations of the employees on the enlargement of their job, which is mostly due to the lack of new tasks assigned after the implementation. This lack of reallocation of tasks after the implementation stresses the importance of the role of the manager to revise the distribution of the tasks within a business unit. Another key takeaway for the managers is the increased skills that are required by the employees within the new situation. Furthermore, the misalignment between the perceived autonomy of the employee and the given autonomy by the manager needs to be addressed. Otherwise, the positive effects of the job enrichment on this sub-dimension will not be realised. The more complex function should be addressed through training on the job or by altering the recruitment requirements of new staff.
5.2.2 Scientific Implications
The theoretical contribution of this paper is the enrichment of the current literature on RPA and the extensions of the basis for further research on this topic. The findings on the influence of RPA on the job design of the involved jobs contribute to the current literature in multiple disciplines. It enlightens the general field of automation by elaborating on a new and revolutionary form of automation, by exposing the effects of RPA on the job design of the involved jobs. Furthermore, it adds to the discussion on the on the impact of the automation and robotic revolution and the influence on the employees as described by Anagoste (2017). Moreover, it adds to the paper Autor (2015) on the benefits and threats, by exposing the benefits of RPA on the involved workforce.
This research consists of a total of ten in-depth interviews to gain insights into the effects of RPA on FTE effort and the job design of the involved jobs. Although these interviews do provide useful information, the outcomes are limited generalizable due to the small research sample. Another limitation of this research is the use of propositions since these propositions do not rely on verifiable data and are therefore only useful to explore the phenomenon of RPA. These propositions need to be transformed into hypotheses and tested by a larger data collection, which can result in outcomes based on statistics to determine whether the hypotheses are valid or not.
The convenience sampling method used in this paper also brings a limitation to the research. Since most of the respondents are clients of KPMG, which result in RPA cases with a similar implementation method, which can influence the effects of the implementation itself. Although other firms are also researched, this influence should be taken into account, and the results have to be interpreted with caution. Furthermore, the managers and employees were selected through convenience sampling which can result in sampling bias. This bias could also be avoided by expanding the research and use a more thorough sampling method.
5.4 Future Research
This research explores the field of RPA and observes the data collected from ten respondents, where the observations indicate that the implementation of RPA results in the enrichment and enlargement of the involved jobs. Although this research extends the current literature on RPA, further quantitative research is needed to find persuasive and statistical evidence for these result and therefore strengthen these findings.
Moreover, the proposed FTE gap was not indicated by the respondents; however, the observations indicate that the lack of proof on this gap can ensue due to the low maturity and