Robotic process automation – research impulses from the BPM 2023 panel discussion

Abstract

Robotic Process Automation is an established technology in organizations. In the last years, it has also received considerable attention in scholarly research with publications, special issues, and academic conferences dedicated to the topic. Given that Robotic Process Automation has now moved beyond the initial hype, we can ask what research should focus on in the future. To address this question, we conducted a panel discussion to discuss its current state and future development. This panel, which took place at the Robotic Process Automation forum at the Business Process Management Conference 2023, included experts from academia and industry, covering strategy consultants, implementers, and tool providers. In this report, we present insights from the panel discussions. We especially focus on three future research directions on Robotic Process Automation that emerged from the panel.

Full text

Open Access
© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits
use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original
author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third
party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the mate‑
rial. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://
creativecommons.org/licenses/by/4.0/.
BRIEF COMMUNICATION
Plattfautetal. Process Science (2024) 1:5
https://doi.org/10.1007/s44311-024-00005-1
Process Science
Robotic process automation – research
impulses fromtheBPM 2023 panel discussion
Ralf Plattfaut1*, Jana‑Rebecca Rehse2*, Caspar Jans3, Matthias Schulte4 and Joost van Wendel de Joode5
Abstract
Robotic Process Automation is an established technology in organizations. In the last
years, it has also received considerable attention in scholarly research with publications,
special issues, and academic conferences dedicated to the topic. Given that Robotic
Process Automation has now moved beyond the initial hype, we can ask what research
should focus on in the future. To address this question, we conducted a panel dis‑
cussion to discuss its current state and future development. This panel, which took
place at the Robotic Process Automation forum at the Business Process Management
Conference 2023, included experts from academia and industry, covering strategy
consultants, implementers, and tool providers. In this report, we present insights
from the panel discussions. We especially focus on three future research directions
on Robotic Process Automation that emerged from the panel.
Keywords: Robotic process automation, Process mining, Cognitive automation, Panel
report
Introduction
Robotic Process Automation (RPA) can be defined as “a technology that allows the devel-
opment of (multiple) computer programs (i.e., bots) that automate rules-based business
processes through the use of GUIs” (Plattfaut and Borghoff 2022). e bots emulate
the actions of a human user, e.g., mouse movements, clicks, or keyboard strokes, which
then are used to interact with underlying core IT systems. As a light-weight automation
technology, RPA provides organizations with the opportunity to automate their busi-
ness processes without changing these underlying IT systems (vanderAalst etal. 2018).
As such, RPA is particularly well-suited for frequently occurring, simple, repetitive, and
manual processes, such as transferring data between systems. However, through the
integration of artificial intelligence (AI) and machine learning techniques, it can also be
applied to automate more complex processes (Engel etal. 2022).
Although its core technology is based on older concepts such as screen-scraping
(Taulli 2020), RPA only became widely popular around 2015 (Kregel etal. 2021). Since
then, it has received considerable attention in industry (Enríquez etal. 2020), academia
(Plattfaut and Borghoff 2022; Syed etal. 2020), and the general public (Kregel etal. 2021).
RPA has attracted several special issues (e.g., Reijers etal. (2021)), has been the topic of
*Correspondence:
Ralf.Plattfaut@icb.uni‑due.de;
rehse@uni‑mannheim.de
1 University of Duisburg‑Essen,
Essen, Germany
2 University of Mannheim,
Mannheim, Germany
3 Celonis Process Management,
Amsterdam, Netherlands
4 viadee AG, Münster, Germany
5 UIPath, Inc., New York, USA

Page 2 of 15
Plattfautetal. Process Science (2024) 1:5
dedicated publication outlets (e.g., Plattfaut and Rehse (2023)) and has been shown to be
perceived as “beyond the hype” (Kregel etal. 2021). Notably, RPA research has been the
focus of multiple research fields, including business process management (BPM) (Plat-
tfaut and Rehse 2023) as well as information systems (IS) (Haase etal. 2024). On the
practical side, RPA has been proven to work reliably (Plattfaut and Rehse 2023), leading
to widespread adoption (Gartner 2021). Following this development, the global market
for RPA solutions has grown considerably and is expected to grow even further in the
following years (Statista 2021). All these aspects are evidence for the degree of maturity
that RPA has achieved.
is remarkable development elicits the questions: What should RPA research focus
on in the future? Which new (methodological) angles are worth to be considered? Sev-
eral literature reviews have already attempted to answer this question (Syed etal. 2020;
Plattfaut and Borghoff 2022). However, as Webster and Watson famously coined, litera-
ture review set out to “analyze the past to prepare for the future” (Webster and Wat-
son 2002). is orientation risks neglecting new and practice-driven perspectives on the
specific phenomenon of RPA, which is unique in the sense that it gathers attention from
both industry and academia at the same time, but potentially with a different focus. To
ensure the relevance of both RPA scholarship and RPA practice and to avoid a further
separation between the two, RPA can benefit from a closer engagement between aca-
demics and practitioners (Vande Ven 2018), particularly when selecting the problems to
be studied and grounding them in reality.
To foster this engagement and initiate a joint discourse on the theory and practice of
RPA (Vande Ven and Johnson 2006), we conducted a panel discussion at the Business
Process Management (BPM) Conference in 2023. e goal of this panel was to bring
together practitioners with different perspectives on RPA and engage them in a discus-
sion with one another as well as the research community present at the conference. Our
panelists stemmed mainly from industry, covering strategy consultants, implement-
ers, and RPA tool providers. ey were joined by two RPA researchers, one serving as
a panelist providing an academic perspective on the topic and one serving as the panel
moderator.
In this paper, we discuss the insights from this panel discussion on the current devel-
opment of RPA and the future directions of the field. As such, the paper serves as a
report on the (practical) aspects of RPA that the panel raised, placed in the context of
current RPA scholarship. It is meant to contribute to the practitioner-engaged scientific
discourse on RPA by providing novel perspectives on the topic. For this purpose, we
firstprovide some background on RPA, contrasting the perspectives from scholarly liter-
ature (reviews) and practice. en, we present the results of our panel discussion. After-
wards, we reflect upon the results in light of the existing literature and present three
calls for further research on RPA, before closing with a short conclusion.
Background: robotic process automation
RPA is an umbrella term for a broad and ever-increasing range of concepts (Czarnecki
and Fettke 2021; vanderAalst etal. 2018). Accordingly, researchers have proposed mul-
tiple different definitions of RPA. ey cover different aspects of RPA, e.g., its focus on
repetitive tasks (Lacity and Willcocks 2016), the emulation of human workers by the

Page 3 of 15
Plattfautetal. Process Science (2024) 1:5
interaction over graphical user interfaces (Penttinen etal. 2018), or the fact that RPA
bots are easily built (Mayer etal. 2018). Literature reviews have set out to consolidate
these definitions (Syed etal. 2020; Plattfaut and Borghoff 2022). We follow their work
and understand RPA as “a technology that allows the development of (multiple) com-
puter programs (i.e., bots) that automate rules-based business processes through the use
of GUIs” (Plattfaut and Borghoff 2022).
Robotic process automation inscholarly literature
Figure 1 outlines the field of RPA as it is currently considered in literature. Very
broadly, scholars see RPA as a novel technology which can be used to (partially) auto-
mate processes within an organizational context. RPA should contribute to a com-
pany’s BPM goals, i.e., to improve the time, cost, quality, or flexibility of their process
execution (Dumas etal. 2018). Accordingly, we can separate existing research into three
streams(Syed etal. 2020; Plattfaut and Borghoff 2022): (1) RPA itself, i.e., the underly-
ing technology, (2) the goals of RPA, i.e., the benefits it brings to a process, and (3) the
organizational context of RPA, i.e., the environment in which it is managed and applied.
Within these streams, which are briefly outlined in the following, researchers mainly
employ behavioral or design-oriented research methods to gain new insights or develop
new solutions for the advancement of RPA.
In the first stream, it is noteworthy that there is not much research on core RPA
technology. e implementation and configuration of rule-based software robots is
considered to be relatively simple and sufficiently mature, such that it is typically left
to industry. Technology-focused RPA research treats it as a black box, on which other
technology is built. For example, RPA bots are orchestrated to achieve large-scale pro-
cess automation (Rizk etal. 2021). Many papers also treat the automated scoping and
configuration of RPA bots, e.g., through the use of process mining (Leno etal. 2021). In
addition, researchers have started to combine core RPA technology with machine learn-
ing features to enable the automation of more cognitively challenging tasks (Reijers etal.
2021).
With regard to the second stream, contemporary research has put RPA in the con-
text of traditional business process management, mainly focusing on achieving instru-
mental goals of improvements regarding time, cost, quality, and flexibility (François etal.
2022). In this context, a lot of effort has been invested into identifying characteristics
of RPA-suitable tasks (Syed etal. 2020) and identifying them in an automated way, e.g.,
from textual process descriptions (Leopold etal. 2018). To this end, prior research used
Fig. 1 The current field of RPA research

Page 4 of 15
Plattfautetal. Process Science (2024) 1:5
predominantly behavioral or design-oriented methods to create both descriptive and
prescriptive insights (Plattfaut and Borghoff 2022).
Most RPA research can be associated with the third stream, i.e., the organiza-
tional context of RPA. is is also supported by two contemporary literature reviews.
Syed etal. (2020) analyzed 125 peer-reviewed and white papers on RPA and derived a
research agenda pertaining to the benefits of RPA, the organizational and technologi-
cal readiness for RPA, the capabilities needed to implement RPA, corresponding imple-
mentation methodologies, and specific technologies and techniques for implementing
RPA. Similarly, Plattfaut and Borghoff (2022) analyzed 82 research articles and devel-
oped research questions covering process suitability, bot design, integration of RPA and
AI, operating model and governance of RPA, and the interplay of humans and RPA. As
such, the contemporary literature reviews mainly call for more research on the way RPA
can be implemented, about the benefits of RPA and how they can be realized, and on the
operation of RPA within an organization.
Robotic process automation inpractice
Particularly as an academic (Vande Ven 2018), it is much easier to get an overview over
scholarly literature on RPA than it is to form an impression of the views of practitioners
on the topic. Nevertheless, we can form such an impression by looking at white papers
(Manyika etal. 2017) or studies (Deloitte 2017) by consulting companies, market reports
by analysts (Gartner 2021), or case study reports (Stenzel etal. 2021). Even when taken
with the necessary grain of salt, such gray literature shows that the conversation about
RPA in practice focuses on similar three streams than the academic discourse (Chugh
etal. 2022): (1) RPA technology, (2) the goals that RPA can achieve, and (3) the organiza-
tional context in which RPA exists. However, within and among these streams, the areas
of focus differ considerably.
e focus of practitioners is on traditional “Class 1” RPA technology, where no arti-
ficial intelligence or machine learning is applied (Chugh etal. 2022). is is mirrored
by the RPA definition in the IEEE Guide for Terms and Concepts in Intelligent Process
Automation, which states that RPA is a “preconfigured software instance that uses busi-
ness rules and predefined activity choreography to complete the autonomous execution
of a combination of processes, activities, transactions, and tasks in one or more unre-
lated software systems to deliver a result or service with human exception management”
(IEEE 2017). Compared to scholarly literature, the focus is more on the concrete imple-
mentation and scaling of the technology (Stenzel etal. 2021). Although there are dis-
cussions on integrating rule-based RPA with more “intelligent” capabilities to widen its
scope of tasks (Manyika etal. 2017), there are few reports on the practical use of such
solutions.
Regarding RPA goals, there is a notable divide between the scholarly and the practi-
cal RPA discourse. Practitioners tend to focus on the tangible, short-term outcomes of
RPA implementation, particularly its ability to improve operational compliance, reduce
costs, and increase productivity (Deloitte 2017). e primary goal is to automate rou-
tine, time-consuming tasks to free up human resources for higher-level work (Manyika
etal. 2017). Additionally, there is an emphasis on achieving faster processing times and
improving accuracy in data handling by reducing human error (Stenzel et al. 2021).

Page 5 of 15
Plattfautetal. Process Science (2024) 1:5
Hence, practitioners are more concerned with immediate performance improvements
and the potential return on investment that RPA can deliver. However, it is unclear how
much of this potential can be realized, given that many companies are still in the process
of RPA implementation (Deloitte 2017).
Compared to the (financial) goals of RPA, practitioners view its organizational context
as a secondary concern (Stenzel etal. 2021). Whereas the broader implications of RPA
for organizational structure, culture, and workforce dynamics are relevant for the aca-
demic discourse, particularly in the information systems discipline, practitioners focus
more on the short-term impact (Deloitte 2017). eir concern is typically around how
RPA fits within existing workflows, whether the organization has the necessary technical
infrastructure, and how employees will adapt to the changes (Deloitte 2017). Resistance
to change, retraining staff, and the integration of RPA with other systems are often prac-
tical challenges, but they are typically seen as hurdles to overcome rather than central
points of strategic focus (Chugh etal. 2022).
A contrast inperspectives
Table1 summarizes the perspectives that academics and practitioners have on RPA,
according to their respective bodies of literature as sketched in this section.
Although we can see some similarities in the respective discourses, we also see that the
weight of the respective areas differs. is might indicate a further divergence between
the fields, which is not beneficial for either side. One way to counteract this divergence is
to enable a joint discourse. is was the main motivation for our panel discussion, which
is described in the next section.
Broadening theeld ofRPA research – results fromthepanel discussion
To get a broader understanding on the current issues in the field of RPA, to assess
whether RPA still deserves additional research, and to direct the focus of this
research, we conducted a panel discussion at the RPA forum of the 2023 BPM Con-
ference. e goal of this panel discussion was to integrate a practical view into the
academic discourse on RPA and bring together the practical and the scholarly per-
spective on the topic. e panel took place within the context of the RPA Forum,
which has been a co-located event at the BPM conference for a few years and typically
brings together large parts of the RPA research community, either as authors or as
Table 1 Contrasting perspectives on RPA
Academia Practice
RPA Technology Defined w.r.t. rule‑based processes Defined as rule‑based software
Considered as black box Considered as black box
Focus on integration of ML Focus on implementation and scale
RPA Goals Process goals Financial goals
Identification of suitable processes Identification of routine tasks
Focus on long‑term effects Focus on short‑term effects
RPA Context Primary concern Secondary concern
Considered on an organizational scale Considered on a project scale
Focus on (theoretical) explanations Focus on (practical) solutions

Page 6 of 15
Plattfautetal. Process Science (2024) 1:5
PC members (Plattfaut and Rehse 2023). As chairs of the 2023 RPA Forum, the two
first authors of this report organized the panel, selected the participants, and outlined
potential questions. One of them served as the panel moderator, the other one served
as a panelist providing an academic perspective on the topic. For the 90-minute panel
discussion, they were joined by four participants from practice: (i) a strategy consult-
ant, (ii) a representative from a tool provider, (iii) a representative from an imple-
mentation company, and (iv) a BPM consultant. ese participants were selected to
achieve a broad view on RPA practice, with employees of different companies provid-
ing complementary perspectives on the topic. All initially selected participants fol-
lowed our invitation.
In the following, we summarize the contents of that discussion, supported by direct
quotes from the practitioner panelists wherever helpful to summarize the contents in
a concise way. Note that the broad topics were previously planned by the two chairs,
but that the discussion was conducted organically. e participants shared their own
views freely, which sometimes opened up topics that were not previously planned.
e Current State of RPA Our panelists agreed with our understanding that RPA is an
established technology. To quote one panelist (tool provider):
“Five years ago, we had a lot of questions around ‘Can we try this [RPA] before
we buy it?’ [...] What you see right now is that we do not get those questions any‑
more. [...] A majority of the business users, but also from IT, now know that it is a
proven technology.”
Apparently, RPA has achieved a certain degree of maturity from a technological per-
spective (Kregel et al. 2021). Also, it has gained acceptance among its users, who
appear to trust its capabilities.
Nevertheless, the panelists also agreed that RPA is still an important topic today with
interesting avenues of development both on the academic and the practitioner side. Fur-
ther research and guidance is needed – especially from the process science community.
During the panel, several phenomena from both ends of the socio-technical spectrum
(Sarker etal. 2019) were discussed.
RPA as One of Many Automation Solutions In essence, RPA is a means to an end: It
supports organizations in reaching a higher degree of automation. However, it is only
one potential solution for this purpose. To quote a panelist (implementor):
“RPA, from my perspective is just one element there in the tool chain of helping
organizations automate things.”
e value of RPA is the highest when no other way of automation is available or (eco-
nomically) viable. However, this principle is not always followed, which causes multi-
ple challenges.
Panelists stressed that in many situations, citizen developers (i.e., non-IT professionals
involved in application development) (Plattfaut 2019; Hallikainen etal. 2018) rely on UI-
based automation where other more stable forms of process automation would be pref-
erable. One panelist shared an anecdote about an early RPA project, where a software

Page 7 of 15
Plattfautetal. Process Science (2024) 1:5
robot was used to extract foreign currency exchange rates from a website and enter them
into an SAP system. is solution worked well until the UI of the website was changed
and the robot began entering wrong data into the system. Because the knowledge about
the implementation had diminished once the process step was automated, it took the
company a week to find and address the problem.
is anecdote underlines the necessity of evaluating the different process automation
options (e.g., RPA or core system automation). Core systems are typically more relia-
ble and more stable than solely UI-based solutions such as RPA. e problem is that
in many situation and for many systems, the required APIs are not readily available.
Moreover, API-based automation is typically more expensive and more complex than
RPA. Whereas API integration requires the involvement of IT specialists, RPA can be
implemented by citizen developers. Additionally, the intricacies of diverse APIs across
different systems and applications pose challenges in achieving seamless connectivity.
is financial and operational complexity can deter organizations from embracing API-
based automation, especially when the perceived benefits do not outweigh the associ-
ated costs.
e Long‑Term Effects of RPA Building on these arguments, the panel argued that RPA
may lead to a potential decline in business cases for core system renewal. When RPA
readily tackles simpler, more immediate business cases, organizations may hesitate to
invest in the overhaul or upgrade of core systems. e allure of automating individual
tasks without significant system-level changes might overshadow the broader need for
strategic core system upgrades. In essence, RPA functions as a bandaid to cover up more
relevant IT problems. To quote one panelist (BPM consultant):
“My personal statement is that the more RPA you have in the process, the worse
the quality of the process is.”
is raises questions about the long-term sustainability and scalability of RPA bots.
Building on this, the panelists underlined the need for insights on how to build migra-
tion-ready RPA bots, a task that has recently been picked up by scholars (Strothmann
and Schulte 2023).
RPA & AI Next to the topic of organizational RPA implementation and use, the panel
also discussed the connection of RPA with other technologies. Current technological
innovations offer considerable potential for reducing this manual effort, increasing both
the scope and the applicability of RPA in practice. e first relevant technological inno-
vation is the use of Artificial Intelligence (AI) techniques, particularly Machine Learning
(ML), which can support RPA in multiple ways. e panelists discussed that AI has the
potential to increase the scope of RPA to also include decision-making and other cogni-
tive tasks. As one panelist (tool provider) put it:
“RPA is the body and AI is the brain.”
is relates to the capabilities of ML models to capture the input factors of human
decisions in processes and train a model to automate these decisions. Combined with
RPA as the process-executing technology, this considerably widens the potentials for
process automation, which is why many RPA suites nowadays include AI components:

Page 8 of 15
Plattfautetal. Process Science (2024) 1:5
e fact that there is [currently] not that much AI on the market, [but] it’s growing
and it needs to be incorporated into your organization. RPA is quite suitable for
this.
is could even be extended by generative AI technology, such as ChatGPT, which also
hold the potential to automate manual creative tasks, such as writing e-mails. However,
as the panelists pointed out, these technologically advanced solutions are often not nec-
essary. In many mass processes, the focus is more on automating rule-based or transac-
tional decision making then on generating new content.
RPA & Process Mining A second technological innovation that is highly relevant for
RPA is process mining (Agostinelli etal. 2022), which holds considerable untapped
potential for the development of more capable RPA bots. e panelists discussed mul-
tiple aspects of integrating process mining and RPA: First, analyzing a process by means
of process mining prior to automating it by means of RPA can lead to a more stream-
lined and improved process, which simplifies the application of RPA. Second, the appli-
cation of process mining to low-level UI log data (Rehse etal. 2024), often called task
mining, can help to implement RPA more quickly and efficiently. ird, process mining
can be extended to more unstructured data, such as e-mails, to enable process automa-
tion in the first place (Khandaker etal. 2024). Finally, process mining can help to find
automation potentials by finding repetitive structured even in large and complex pro-
cesses (Leno etal. 2021).
e participants stressed that the potentials of technological innovations, such as AI
and process mining, are particularly relevant in non-standardized processes, where the
definition and scoping of processes is a necessary prerequisite for RPA. For these sys-
tems, we need to establish the conceptual and technological boundaries of a process
before even thinking about automating it. To quote one participant (implementor):
“In a happy SAP world with no customization in place, [RPA] works pretty fine.
However, in all the organizations we are consulting, we have a very heterogeneous IT
landscape with individual systems that maybe not even have something like a task
identifier or case identifier.”
is shows that for RPA to be applied outside of standardized, well-scoped processes,
academics need to work on better leveraging the capacities of process mining and auto-
mated process analysis.
Knowledge Loss due to RPA On top of these technological themes, the discussion also
covered the impact on people. One critical concern that emerged during the panel dis-
cussion was the potential for knowledge loss within organizations due to the widespread
adoption of RPA. As organizations increasingly automate tasks, there is a risk that pro-
cess knowledge may be eroded at both individual and organizational levels and corre-
sponding skills decay (Mirispelakotuwa etal. 2023; Vu etal. 2023). With the delegation
of routine tasks to RPA bots, employees may gradually lose the hands-on experience and
expertise required to manually execute processes. is could be particularly problematic
during scenarios such as system updates, upgrades, or unforeseen circumstances that
necessitate a temporary return to manual operations. Individuals who have become reli-
ant on automated processes may struggle to recall the intricacies of manual execution,

Page 9 of 15
Plattfautetal. Process Science (2024) 1:5
leading to delays and potential errors. As one panelist recalled from an insurance com-
pany they consulted:
“A couple of years after the automation the process knowledge is lost.”
On an organizational level, the cumulative impact of individual knowledge erosion poses
a broader challenge. As processes become more automated, the institutional memory
associated with manual workflows may fade. is becomes a significant risk when
organizations need to revisit manual execution, either due to system changes or stra-
tegic shifts. e loss of organizational process knowledge could hinder adaptability and
increase the learning curve when reverting to manual operations.
However, the single RPA bots are typically developed and visible in a Graphical User
Interface. As such, reverse-engineering the process and the related process knowledge
from the RPA bot is possible. One panelists argued that “there are people using RPA just
to describe their own process”.
e Ethics of RPA Lastly, the panel emphasized the paramount importance of delving
into the ethical dimensions associated with the deployment of automated processes. is
opens up a broader field encompassing the ethical implications of process automation,
transcending the immediate concerns of RPA implementation. As organizations increas-
ingly leverage RPA to streamline operations and enhance efficiency, the ethical implica-
tions of these technological advancements come to the forefront. e panel highlighted
the need for comprehensive research in this field, considering the potential impact on
societal structures. is could start with further insights on the “human in the loop in
all that automation”, as one panelist put it. However, this includes the need to explore
the ethical considerations surrounding job displacement and the potential impact on the
workforce (Willcocks 2020). Understanding societal attitudes and concerns can inform
the development of ethical guidelines that align with public values.
As RPA continues to reshape organizational processes, understanding and addressing
the ethical implications of process automation is paramount. Comprehensive research in
this field will contribute to the development of ethical frameworks that guide responsi-
ble RPA adoption, ensuring a harmonious integration of technology with societal values
and well-being.
Three calls forfuture research
e topics discussed on the panel broaden the field of RPA research. As indicated above,
prior research mostly focused on the implementation of RPA itself and its effect on tra-
ditional dimensions such as time, cost, quality, and flexibility (Syed etal. 2020; Plattfaut
and Borghoff 2022; François etal. 2022). Building on this, our panel suggests that RPA
research should also focus on the social effects of RPA regarding the society, the organi-
zation, and the individual. Moreover, RPA research can also discuss the integration of
RPA with other technologies such as AI, ML, or process mining. As an example, the pan-
elists opened up interesting questions on the integration of RPA and generative artifi-
cial intelligence. Last, the insights gained from the panel also suggest to consider new
methodological angles to RPA research. Contemporary research predominantly relies

Page 10 of 15
Plattfautetal. Process Science (2024) 1:5
on behavioral or design-oriented methods to create descriptive or prescriptive insights.
Next to this, researchers should also take a critical perspective (Myers and Klein 2011)
on the current practices associated with RPA to reveal opportunities for emancipation
of individuals (Young 2023).Our three calls for future research are summarized in Fig.2
and elaborated in the following.
The social aspects ofRPA
The widespread use of RPA has led to profound implications for organizations. Pro-
cesses could be improved regarding time, cost, quality - and sometimes flexibility
(François etal. 2022). However, next to these instrumental values that RPA targeted,
RPA has also considerable social effects, which can be on an individual, organi-
zational, or societal level. The panel in particular discussed the effects of RPA on
knowledge loss, both at the individual and at the organizational level. Due to the
increasing automation of business processes, individuals and organizations lose the
knowledge to execute these processes manually. This opens up interesting ques-
tions of knowledge management and preservation. Research has recently begun to
tackle these questions (Mirispelakotuwa etal. 2023; Eulerich etal. 2024), but further
research appears to be still necessary.
More broadly speaking, the application of RPA has an overall effect on the over-
arching socio-technical system. which can be understood in terms of task, human,
technology, and structure (Heinzl etal. 2024). The deployment of RPA can be con-
sidered a socio-technical change which impacts all four components. Regarding the
technology, new tools and methods are applied to develop the envisioned bot. More-
over, RPA interrelates with one or several existing application systems and is inte-
grated into the organization’s technological infrastructure. Regarding the task, RPA
alters how and by whom the tasks of a process is executed. This might impact not
only the performance level of the execution, but also the execution itself, especially
if the process is optimized towards RPA. Regarding the humans, project participants
and stakeholders influence RPA development, operation, and adoption through their
characteristics, interests, and expectations. In addition, their jobs and tasks might
also get impacted by RPA implementation, which hints towards reciprocal effects
on their attitude towards the robots. Regarding the structure, an RPA initiative is
embedded into an organizational context. Institutional structures, strategies, and
guidelines form the humans’ choices and behavior, thereby affecting project organi-
zation and decision-making.
Fig. 2 Broadening the field of RPA research

Page 11 of 15
Plattfautetal. Process Science (2024) 1:5
The characteristics of the individual socio-technical systems, their components,
and their dynamics determine how successful an RPA implementation will be (Lyyt-
inen and Newman 2008; Wallace etal. 2021). Hence, some systems may be more
suitable for the application of RPA than others, and it is important to assess their
suitability holistically by including all four components into the assessment, to
derive the most suitable use cases for RPA within an organization (Hofmann etal.
2020) and to better understand the social aspects of RPA (François et al. 2022).
Methodologically, authors can rely on empirical research methods (Easterbrook
etal. 2008), especially from the qualitative tradition (Davison 2023).
The integration ofRPA withmachine learning and(generative) articial intelligence
e combination of standard RPA with emerging technologies from the field of artificial
intelligence opens up remarkable opportunities in terms of the scope of RPA, meaning
the types of tasks that the robots can conduct, and its reliability, meaning the ability of
the robots to handle exceptions. Accordingly, this integration of RPA with other technol-
ogies can increase the effects that RPA has, both with regard to process improvement,
but also with regard to the social effects sketched above. is integration of RPA with
other technologies for the sake of automating more cognitive tasks has already been dis-
cussed in literature (Mendling etal. 2018; Viehhauser 2020). Still, the panel highlighted
the importance of further developing and strengthening this line of research, in parts
because RPA can be considered as a vehicle to integrate AI capabilities into processes
without requiring a fundamental re-design of the underlying IT systems. In this sense,
the benefits of RPA as a light-weight, democratized automation techniques remain the
same, but the technological scope and power become much larger.
e panel focused on the integration of RPA with two technologies in particular: gen-
erative AI and process mining. e latter has already been discussed at various places
in literature (El-Gharib and Amyot 2023). Nevertheless, the panel stressed its contin-
ued relevance, envisioning a future development where the enhanced capabilities of RPA
moves it from a simple commodity to a central technology for process automation. is
means that it has direct impact on the management of those processes, making the use
of process mining even more important: In a cyclic fashion, the process is executed by
means of RPA and analyzed by means of process mining. rough the continued identi-
fication of improvement potentials and the generation of new data that can be used for
training (cognitive) RPA bots, this tight integration paves the way towards a continued
automated process improvement.
e second technology that should be closer integrated with RPA is generative AI, par-
ticularly in the form of large language models (LLM). Different from process mining, this
topic has so far not been widely discussed in research, mainly because powerful LLMs
like GPT-4 have only been available to the public for a short amount of time. Researchers
have investigated the capabilities of LLMs to configure software robots based on event
logs (FaniSani et al. 2023). However, the potentials of generative AI are much larger
when used to provide input for the robot, e.g., by extracting the relevant information out
of a lengthy document, or to collaborate with the robot to generate output, e.g., by writ-
ing e-mails (Haase etal. 2024). Hence, generative AI provides many interesting avenues
for future research to advance the development of cognitive RPA.

Page 12 of 15
Plattfautetal. Process Science (2024) 1:5
Work on this future research area could rely on more practice-integrated meth-
ods such as action research (Baskerville and Pries-Heje 1999), design science research
(Hevner etal. 2004), action design research (Sein etal. 2011), or clinical research (Bask-
erville etal. 2023).
Critical research perspectives onRPA
Contemporary research on RPA has focused on behavioral and design-oriented meth-
ods (Plattfaut and Borghoff 2022). Scholars studied the effect of RPA on organizations
and derived implementation models for RPA or associated critical success factors.
However, ethical questions come more and more into focus. An increasing automa-
tion will lead to both psychological (Haase etal. 2024) and societal (Manyika etal.
2017) effects that require further research. However, RPA with its seemingly light-
weight approach can also empower users through the development of automation
solutions. As such, it is one way to democratize IT development and empower or
emancipate employees in the organization to shape their own IT application portfolio
(Godefroid etal. 2024).
Analyzing such value-laden questions requires taking a critical perspective (Myers
and Klein 2011; Young 2023). “Critical research in information systems is concerned
with social issues such as freedom, power, social control, and values with respect to
the development, use, and impact of information technology” (Myers and Klein 2011).
It has been used to tackle ethical questions, e.g., regarding the effects of COVID-19
tracing apps (Rowe etal. 2020), or to develop guidelines to create empowering gov-
ernance structures (Zubler etal. 2024).
From a methodology perspective, a plethora of critical research methods in IS research
have been explored recently, e.g., in the form of emancipatory design science (Young
2023). Extending the portfolio of methods used in RPA research to also include these
critical research methods will open up the field to new and exciting insights.
Conclusion
Although RPA is now an established field, there is still ample need for future research.
e panelists agreed that RPA is relevant for both research and practice. However, RPA
is constantly evolving as it is combined and merges with more and more adjacent con-
cepts such as AI, ML, or process mining. As such, future research can, e.g., consider the
social aspects of RPA, study the integration of RPA with other technological advances,
or take critical research perspectives to evaluate the broader ethical implications. For
those future research areas, potential methodological approaches are discussed in this
report. As such, the panel and the related report provided answers to the questions on
future research areas and related new methodological angles of study.
While we set out to accompany existing literature reviews on RPA with a more prac-
tice-driven perspective, we need to acknowledge that these practice-driven perspectives
of course root in the experiences of the panelists. While we opted for a broad range of
different perspectives (i.e., tool providers, consultants, BPM experts), our results are still
limited by the selection of panelists and their idiosyncratic experience-based opinions.

Page 13 of 15
Plattfautetal. Process Science (2024) 1:5
In conclusion, our panel not only underlined the importance of the topic but also
outlined areas for future research. We look forward to this research, especially from
the process science community.
Authors’ contributions
R.P. and J.R. prepared the panel and wrote the report. R.P. moderated the panel. J.R., C.J., M.S., and J.W. participated in the
panel. C.J., M.S., and J.W. reviewed the report.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
Not applicable.
Competing interests
The authors declare no competing interests.
Received: 5 February 2024 Accepted: 20 October 2024
References
Agostinelli S, Marrella A, Abb L, et al (2022) Mastering robotic process automation with process mining. In: Business
Process Management. Springer, Cham, p 47–53
Baskerville R, Pries‑Heje J (1999) Grounded action research: a method for understanding it in practice. Account
Manag Inf Technol 9(1):1–23. https:// doi. org/ 10. 1016/ S0959‑ 8022(98) 00017‑4
Baskerville R, Vom Brocke J, Mathiassen L et al (2023) Clinical research from information systems practice. Eur J Inf
Syst 32(1):1–9. https:// doi. org/ 10. 1080/ 09600 85X. 2022. 21260 30
Chugh R, Macht S, Hossain R (2022) Robotic process automation: a review of organizational grey literature. Int J Inf
Syst Proj Manag 10(1):5–26
Czarnecki C, Fettke P (2021) Robotic process automation: Positioning, structuring, and framing the work. In: Czar‑
necki C, Fettke P (eds) Robotic Process Automation. De Gruyter, Oldenbourg, pp 3–24
Davison RM (ed) (2023) Handbook of Qualitative Research Methods for Information Systems. Edward Elgar Publish‑
ing, Cheltenham. https:// doi. org/ 10. 4337/ 97818 02205 398
Deloitte (2017) The robots are ready. are you? Untapped advantage in your digital workforce. https:// www2. deloi
tte. com/ conte nt/ dam/ Deloi tte/ tr/ Docum ents/ techn ology/ deloi tte‑ robots‑ are‑ ready. pdf. Accessed 18 Sept 2024
Dumas M, La Rosa M, Mendling J et al (2018) Fundamentals of business process management. Springer, Berlin
Easterbrook S, Singer J, Storey MA, et al (2008) Selecting Empirical Methods for Software Engineering Research. In:
Shull F, Singer J, Sjøberg DIK (eds) Guide to Advanced Empirical Software Engineering. Springer London, Lon‑
don, pp 285–311. https:// doi. org/ 10. 1007/ 978‑1‑ 84800‑ 044‑5_ 11
El‑Gharib NM, Amyot D (2023) Robotic process automation using process mining–a systematic literature review. Data
Knowl Eng 148:102229
Engel C, Ebel P, Leimeister JM (2022) Cognitive automation. Electron Mark 32(1):339–350
Enríquez JG, Jiménez‑Ramírez A, Domínguez‑Mayo FJ et al (2020) Robotic process automation: a scientific and indus‑
trial systematic mapping study. IEEE Access 8:39113–39129
Eulerich M, Waddoups N, Wagener M et al (2024) The Dark Side of Robotic Process Automation (RPA): Understanding
Risks and Challenges with RPA. Account Horiz 38(2):143–152
Fani Sani M, Sroka M, Burattin A (2023) Llms and process mining: Challenges in rpa: Task grouping, labelling and con‑
nector recommendation. In: Process Mining Workshops. Springer, Cham, p 379–391
François PA, Borghoff V, Plattfaut R et al (2022) Why companies use RPA: a critical reflection of goals. In: Di Ciccio C,
Dijkman R, Del Río Ortega A et al (eds) Business Process Management, vol 13420. Springer International Publish‑
ing, Cham, pp 399–417
Gartner (2021) Magic quadrant report for robotic process automation. https:// www. gartn er. com/ en/ docum ents/
39880 21/ magic‑ quadr ant‑ for‑ robot ic‑ proce ss‑ autom ation. Accessed 18 Sept 2024
Godefroid ME, Borghoff V, Plattfaut R, et al (2024) Structural power imbalances in global organisations: analysing it
governance from a postcolonial perspective. Eur J Inf Syst 1–22
Haase J, Kremser W, Leopold H, et al (2024) Interdisciplinary directions for researching the effects of robotic process
automation and large language models on business processes. Commun Assoc Inf Syst 54:54
Hallikainen P, Bekkhus R, Pan SL (2018) How OpusCapita Used Internal RPA Capabilities to Offer Services to Clients.
MIS Q Exec 17(1):41–52

Page 14 of 15
Plattfautetal. Process Science (2024) 1:5
Heinzl A, Mädche A, Riedl R (2024) Wirtschaftsinformatik: Einführung und grundlegung. Springer, Berlin Heidelberg
Hevner AR, March ST, Park J et al (2004) Design science in information systems research. MIS Q 28(1):75–105
Hofmann P, Samp C, Urbach N (2020) Robotic process automation. Electron Mark 30(1):99–106
IEEE (2017) Ieee guide for terms and concepts in intelligent process automation. https:// ieeex plore. ieee. org/ docum
ent/ 80706 71. Accessed 18 Sept 2024
Khandaker F, Senderovich A, Zhao J et al (2024) Transformer models for mining intents and predicting activities from
emails in knowledge‑intensive processes. Eng Appl Artif Intell 128:107450
Kregel I, Koch J, Plattfaut R (2021) Beyond the hype: Robotic process automation’s public perception over time. J
Organ Comput Electron Commer 31(2):130–150
Lacity MC, Willcocks LP (2016) Robotic process automation at telefónica o2. MIS Q Exec 15(1):21–35
Leno V, Polyvyanyy A, Dumas M et al (2021) Robotic process mining: Vision and challenges. Bus Inf Syst Eng
63(3):301–314
Leopold H, van der Aa H, Reijers H (2018) Identifying candidate tasks for robotic process automation in textual pro‑
cess descriptions. In: Enterprise, Business‑Process and Information Systems Modeling. Springer, Cham, p 67–81
Lyytinen K, Newman M (2008) Explaining information systems change: a punctuated socio‑technical change model.
Eur J Inf Syst 17(6):589–613
Manyika J, Lund S, Chui M, et al (2017) Jobs lost, jobs gained: workforce transitions in a time of automation. McKinsey
Global Institute, Washington, DC. https:// www. mckin sey. com/ featu red‑ insig hts/ future‑ of‑ work/ jobs‑ lost‑ jobs‑
gained‑ what‑ the‑ future‑ of‑ work‑ will‑ mean‑ for‑ jobs‑ skills‑ and‑ wages. Accessed 18 Sept 2024
Mayer J, Esswein M, Razaqi T, et al (2018) Zero‑quartile benchmarking ‑ a forward‑looking prioritization of digital tech‑
nologies for a company’s transformation. In: Pries‑Heje J, Ram S, Rosemann M (eds) Proceedings of the International
Conference on Information Systems. Association for Information Systems
Mendling J, Decker G, Hull R et al (2018) How do Machine Learning, Robotic Process Automation, and Blockchains Affect
the Human Factor in Business Process Management? Commun Assoc Inf Syst 43:297–320
Mirispelakotuwa I, Syed R, Wynn MT (2023) Is RPA Causing Process Knowledge Loss? Insights from RPA Experts. In: Köpke
J, López‑Pintado O, Plattfaut R et al (eds) Business Process Management: Blockchain, Robotic Process Automation
and Educators Forum, vol 491. Springer Nature Switzerland, Cham, pp 73–88
Myers Klein (2011) A set of principles for conducting critical research in information systems. MIS Q 35(1):17
Penttinen E, Kasslin H, Asatiani A (2018) How to choose between robotic process automation and back‑end system
automation? In: Bednar P, Frank U, Kautz K (eds) Proceedings of the 26th European Conference on Information Sys‑
tems. Aston University
Plattfaut R (2019) Robotic process automation ‑ process optimization on steroids? In: Krcmar H, Fedorowicz J, Boh WF,
et al (eds) Proceedings of the 40th International Conference on Information Systems. Association for Information
Systems
Plattfaut R, Borghoff V (2022) Robotic process automation: A literature‑based research agenda. J Inf Syst 36(2):173–191
Plattfaut R, Rehse J (2023) Preface of the Proceedings of the RPA Forum. In: Köpke J, López‑Pintado O, Plattfaut R et al
(eds) Business Process Management: Blockchain, Robotic Process Automation and Educators Forum. Springer
Nature Switzerland, Cham, pp 67–71
Rehse JR, Abb L, Berg G, et al (2024) User behavior mining. Bus Inf Syst Eng 1–18
Reijers HA, Wynn MT, van de Weerd I, et al (2021) Robotic process automation ‑ special issue in computers in industry
Rizk Y, Chakraborti T, Isahagian V et al (2021) Towards end‑to‑end business process automation. In: Czarnecki C, Fettke P
(eds) Robotic Process Automation. De Gruyter, Oldenbourg, pp 155–168
Rowe F, Ngwenyama O, Richet JL (2020) Contact‑tracing apps and alienation in the age of covid‑19. Eur J Inf Syst
29(5):545–562
Sarker S, Chatterjee S, Xiao X et al (2019) The sociotechnical axis of cohesion for the is discipline: Its historical legacy and
its continued relevance. MIS Q 43(3):695–719
Sein MK, Henfridsson O, Purao S et al (2011) Action design research. MIS Q 35(1):37–56. https:// doi. org/ 10. 2307/ 23043 488
Statista (2021) Spending on robotic process automation (rpa) software worldwide from 2020 to 2032. https:// www. stati
sta. com/ stati stics/ 13093 84/ world wide‑ rpa‑ softw are‑ market‑ size/. Accessed 18 Sept 2024
Stenzel A, Ritschel K, Stummer C (2021) The broad use of RPA based on three practical cases. In: Czarnecki C, Fettke P
(eds) Robotic Process Automation. De Gruyter, Oldenbourg, pp 377–391
Strothmann A, Schulte M (2023) Migrating from RPA to Backend Automation: An Exploratory Study. In: Köpke J, López‑
Pintado O, Plattfaut R et al (eds) Business Process Management: Blockchain, Robotic Process Automation and Educa‑
tors Forum, vol 491. Springer Nature Switzerland, Cham, pp 149–164
Syed R, Suriadi S, Adams M et al (2020) Robotic process automation: Contemporary themes and challenges. Comput Ind
115:103162
Taulli T (2020) The Robotic Process Automation Handbook: A Guide to Implementing RPA Systems. Apress L. P, Berkeley
van der Aalst W, Bichler M, Heinzl A (2018) Robotic process automation. Bus Inf Syst Eng 60(4):269–272
Van de Ven AH (2018) Academic‑practitioner engaged scholarship. Inf Organ 28(1):37–43
Van de Ven AH, Johnson PE (2006) Knowledge for theory and practice. Acad Manag Rev 31(4):802–821
Viehhauser J (2020) Is robotic process automation becoming intelligent? Early evidence of influences of artificial intel‑
ligence on robotic process automation. In: Business Process Management: Blockchain and Robotic Process Automa‑
tion Forum. Springer, pp 101–115
Vu H, Haase J, Leopold H et al (2023) Towards a theory on process automation effects. In: Di Francescomarino C, Burattin
A, Janiesch C et al (eds) Business Process Management Forum, Lecture Notes in Business Information Processing, vol
490. Springer Nature Switzerland, Cham, pp 285–301

Page 15 of 15
Plattfautetal. Process Science (2024) 1:5
Wallace E, Waizenegger L, Doolin B (2021) Opening the black box: exploring the socio‑technical dynamics and key princi‑
ples of RPA implementation projects. In: ACIS 2021 Proceedings, 86. Association for Information Systems
Webster J, Watson RT (2002) Analyzing the past to prepare for the future: Writing a literature review. MIS Q 26(2):xiii–xxiii
Willcocks L (2020) Robo‑Apocalypse cancelled? Reframing the automation and future of work debate. J Inf Technol
35(4):286–302
Young A (2023) Critical is research. In: Davison RM (ed) Handbook of Qualitative Research Methods for Information
Systems. Edward Elgar Publishing, Cheltenham, pp 163–181
Zubler ME, Plattfaut R, Niehaves B (2024) Decolonizing IT governance in international non–governmental organisations:
An Ubuntu approach. Inf Syst J
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.