Daniel Toll*, Fredrik Söderström** Linköping University , SE-581 83 Linköping , Sweden 365 369

Robotic Process Automation (RPA) is being adapted by public sector organizations as a means to solve challenges yet new problems and challenges arise. One reason for this may be a mismatch between how RPA is portrayed and what it turns out to be. This paper covers the first analysis in a study that compares the portrayal of RPA by vendors by that of public sector.

 Robotic Process Automation RPA Service Automation Digitalization Digital Transformation Public Sector ICT 1. Introduction

Robotic Process Automation (RPA) is software automation of repetitive tasks. RPA has received significant interest due to its potential to solve some of the current challenges of public sector organizations. The vendors have grand visions and high hopes that RPA solutions will have an alhumans. However, there are challenges ahead and new problems arising when adopting and implementing RPA. One potential reason may be due to how RPA is portrayed and what it is received to be. We are looking into the differences of how RPA is portrayed and received by the vendors and the public sector. So far we have performed an analysis of the vendor portrayal of RPA, which we present in this paper.

 2. Method

We have conducted a content analysis (Krippendorf, 2004) inspired by grounded theory (Glaser and Strauss, 1967) . The study is qualitative and interpretative (Walsham, 1995; 1993) . The data used is content from websites of the five largest RPA vendors (Gartner, 2019) ; a ranking based on estimated market share in 2018. These websites describe what RPA is. Prior to analysis the text from these websites were extracted. The steps performed during analysis were: (1) Open coding, where each sentence were coded with its messasge(message) and type of message (category) as shown in the example in Table 1, (2) Cleaning of categories, where categories were consolidated or removed, groups were also formed to categorize categories, (3) Clustering of messages, similar messages were clustered together, and (4) Filtering, where only clusters based on messages from at least three of the five vendors were kept for rigor. the technology that allows anyone today to configure computer

3. Results Description of RPA (describing statement) Feature (a trait of the technology) Operation (information about how it works) Capability (what the technology is capable of) Purpose (the purpose of the technology)

We present the results in Figure 1 and Figure 2. Figure 1 shows the relative distribution of clusters within their categories and groups. Figure 2 shows the groups, categories and clusters. In total, the results are based on 428 coded rows. In Figure 1, the inner circle represents the groups and the outer circle the categories. The size of the categories is determined by the number of clusters for the respective category. A group contains categories, a category contains clusters, and a cluster is a collection of similar messages.

 4. Conclusions So Far

Emphasizes arguments of acquiring RPA, followed by RPA usage and to a lesser degree the definition of RPA. The overall portrayal is positive but, in some parts, vague. The highly optimistic perception of RPA may cause the requirements and efforts needed to be underestimated.

Is not fully comprehensive, as there are areas of relevance that are covered minimally or not at all. This is unsurprising since the data could be considered marketing material. The absence of certain dimensions may however lead to underestimating requirements and efforts needed for successful RPA implementation and usage.

Is associated or related to AI, which in turn could risk RPA being interpreted as a smarter concept than it is. This can lead to confusion as to what is what and for which purposes the respective technologies can be used.

 5. Continuation

We plan to continue this study by including empirical data from the public sector, to cover both of these perspectives. This is something we already have some data for. We will then compare the two portrayals to spot differences, problematize about the reasons for these differences and discuss possible implications.

About the Authors Daniel Toll Daniel Toll has a background in cognitive science and information systems and is currently working towards his PhD in information systems at Linköping University, Sweden. His research is focused on how the use of artificial intelligence and automation technologies in public sector organizations affect society. Fredrik Söderström

 Gartner. ( 2019 ). Gartner Says Worldwide Robotic Process Automation Software Market Grew 63 % in 2018 . Retrieved from https://www.gartner.com/en/newsroom/press-releases/2019-06-24 -gartner-saysworldwide-robotic-process-automation-sof Glaser , B. , & Strauss , A. ( 1967 ). The discovery grounded theory: strategies for qualitative research . Chicago: Aldin. Krippendorff , K. ( 2004 ). Content analysis: An introduction to its methodology (2nd ed .). Thousand Oaks: Sage Publications. Walsham , G. ( 1993 ). Interpreting information systems in organizations: Wiley Chichester . Walsham , G. ( 1995 ). Interpretive case studies in IS research: nature and method . European Journal of Information Systems , 4 ( 2 ), 74 - 81 .