Factors influencing robotic process 
automation adoption in the South 

African insurance industry. 

. .”  

Namdipha Kunene 

Student number: 1405197 

 

A research thesis submitted to the Faculty of Commerce, Law and 

Management, University of the Witwatersrand, in partial fulfilment of the 

requirements for the degree of Master of Management in the field of  

Digital Business 

 

Johannesburg, 2023 

 

 

 



ii 

ABSTRACT 

Robotic process automation (RPA) is fast becoming a key instrument in the digital 

transformation journey of the insurance market. This software technology 

presents a wide range of benefits to an insurance organisation, from driving 

operational efficiencies to improving customer experience. However, despite the 

vast use cases of RPA evidenced in the global insurance market, the South 

African insurance industry has proven to be a laggard in adopting RPA. In 

understanding the slow uptake of RPA by SA insurers, it was important to gain 

insights on RPA adoption from the perspective of RPA vendors as well as South 

African insurance professionals. A qualitative exploratory study was conducted 

to investigate the antecedents and factors influencing RPA adoption within the 

South African insurance industry. 

This study is based on the exploration of multiple ICT adoption frameworks 

existing in literature, such as user acceptance models used at an individual level 

and diffusion theories used at an organisational level adoption. A conceptual 

framework integrating the technology-organisation-environment (TOE) 

framework, diffusion of innovation (DOI) and technology acceptance model 

(TAM) was created and used to identify the factors. The empirical findings of this 

study were based on twelve (12) semi-structured interviews conducted with 

senior managers within the insurance industry and in the RPA vendor space. 

The study reveals that relative advantage, complexity, compatibility, 

management support, competition pressures and vendor support were perceived 

to have a positive influence on the acceptance and adoption of RPA. The study 

also suggests that strategy and government regulations have a non-significant 

influence on the adoption of RPA, whereas skills in the organisational context 

were perceived as a negative influence. Interestingly, two new factors emerged 

from the thematic analysis conducted: while perceived costs were viewed as a 

negative influence, change management was perceived as a significantly positive 

influence on the adoption of RPA within the SA insurance industry. The study 

provides recommendations to leaders to ensure a seamless RPA adoption 

process.   



iii 

KEYWORDS 

Robotic Process Automation, Technology Adoption, Insurance, Automation 



iv 

DECLARATION 

I, Namdipha Kunene (1405197), declare that this research report is my own work 

except as indicated in the references and acknowledgements. It is submitted in 

partial fulfilment of the requirements for the degree of Master of Management in 

the field of Digital Business at the University of the Witwatersrand, Johannesburg. 

It has not been submitted before for any degree or examination in this or any 

other university. 

Name:  Namdipha Kunene     Signature:  

 

Signed at ………Johannesburg…………………………………………… 

 

On the ………10th ………… day of …February…………………2023. 

 



v 

DEDICATION  

 

This thesis is wholeheartedly dedicated to my late great-grandfather.  

Ndoda Ernest Kunene (1932 - 2022) 

Mntimande, Madonsela, Lubambo lunye, zingaba mbil’ufuz’ekhabonina. 

Ngiyabonga Mkhulu. 

For your endless love, encouragement, and support.  

 

 

 

 

 

 

 

 

 

 

 

 

 



vi 

ACKNOWLEDGEMENTS 

I would like to extend my gratitude to the following people:  

• To my mother, Nompumelelo Kunene, for praying with me and believing 

in me throughout this journey.  

• To my colleagues, for accommodating my busy schedule for the past two 

years and supporting me.  

• To my supervisor, Dr Gregory Lee, for guiding and challenging my thought 

process throughout this study. You have positively influenced the growth 

of my research skills.  

• Last, but not least, I want to thank ME: for doing all this hard work, not 

giving up, and finishing off what I had started.  

“With man this is impossible, but with God all things are possible.” – 

Matthew 19:26 

  



vii 

TABLE OF CONTENTS 

ABSTRACT ....................................................................................................... ii 

DECLARATION ................................................................................................ iv 

DEDICATION ..................................................................................................... v 

ACKNOWLEDGEMENTS ................................................................................. vi 

LIST OF TABLES .............................................................................................. x 

LIST OF FIGURES .......................................................................................... xii 

LIST OF ACRONYMS .................................................................................... xiii 

CHAPTER 1. INTRODUCTION ......................................................................... 1 
1.1 BACKGROUND OF THE STUDY ................................................................................................. 1 
1.2 RESEARCH PROBLEM ............................................................................................................. 2 
1.3 STATEMENT OF PURPOSE ...................................................................................................... 3 
1.4 RESEARCH QUESTIONS .......................................................................................................... 3 
1.5 RATIONALE ........................................................................................................................... 4 
1.6 DELIMITATIONS OF THE STUDY................................................................................................ 4 
1.7 DEFINITION OF TERMS ............................................................................................................ 5 
1.8 ASSUMPTIONS ....................................................................................................................... 6 
1.9 CHAPTER OUTLINE ................................................................................................................. 7 

CHAPTER 2. LITERATURE REVIEW AND THEORETICAL FRAMEWORK .... 9 
2.1 INTRODUCTION ...................................................................................................................... 9 
2.2 SOUTH AFRICAN INSURANCE INDUSTRY .................................................................................. 9 
2.3 ADOPTION OF TECHNOLOGY IN THE SA INSURANCE INDUSTRY ............................................... 10 
2.4 ROBOTIC PROCESS AUTOMATION (RPA) ............................................................................... 11 
2.5 TYPES OF RPA, BENEFITS, AND LIMITATIONS ........................................................................ 12 

2.5.1 Types of RPA ........................................................................................................ 12 
2.5.2 RPA benefits ......................................................................................................... 13 
2.5.3 Limitations of RPA ................................................................................................ 14 

2.6 RPA IN INSURANCE ............................................................................................................. 15 
2.6.1 Antecedents to RPA adoption .............................................................................. 16 

2.7 ANALYTICAL FRAMEWORK ........................................................................................... 19 
2.7.1 Technology-organisation-environmental framework (TOE).................................. 19 
2.7.2 Technology acceptance model (TAM) .................................................................. 22 
2.7.3 Diffusion of innovations theory (DOI).................................................................... 24 
2.7.4 Integration of the adoption models ....................................................................... 26 

2.8 CONCEPTUAL FRAMEWORK .................................................................................................. 27 
2.8.1 Adoption process .................................................................................................. 29 
2.8.2 Technological factors ............................................................................................ 29 
2.8.3 Organisational factors ........................................................................................... 31 
2.8.4 Environmental factors ........................................................................................... 32 
2.8.5 Perceived ease of use (PEOU) and perceived usefulness (PU) of RPA ............. 34 

2.9 SUMMARY OF CHAPTER 2 .................................................................................................... 34 

CHAPTER 3. RESEARCH METHODOLOGY .................................................. 35 
3.1 RESEARCH METHODOLOGY AND DESIGN ............................................................................... 35 
3.2 UNIT OF ANALYSIS ............................................................................................................... 37 
3.3 POPULATION ....................................................................................................................... 38 
3.4 SAMPLING METHODS AND SIZES ........................................................................................... 38 
3.5 DATA COLLECTION PROCESS ................................................................................................ 41 
3.6 RESEARCH INSTRUMENT ...................................................................................................... 42 

3.6.1 Semi-structured interviews ................................................................................... 42 



viii 

3.6.2 Documents ............................................................................................................ 43 
3.7 DATA ANALYSIS STRATEGY ................................................................................................... 44 
3.8 QUALITY ASSURANCE .......................................................................................................... 45 

3.8.1 Transferability ....................................................................................................... 45 
3.8.2 Credibility .............................................................................................................. 45 
3.8.3 Dependability ........................................................................................................ 46 
3.8.4 Confirmability ........................................................................................................ 46 

3.9 ETHICAL CONSIDERATIONS ................................................................................................... 47 
3.9.1 Voluntary participation .......................................................................................... 47 
3.9.2 Consent ................................................................................................................ 47 
3.9.3 Confidentiality and anonymity ............................................................................... 47 

3.10 LIMITATIONS ........................................................................................................................ 48 

CHAPTER 4. RESEARCH FINDINGS AND DATA ANALYSIS ....................... 49 
4.1 INTRODUCTION .................................................................................................................... 49 
4.2 THEMATIC ANALYSIS ............................................................................................................ 49 

4.2.1 Coding .................................................................................................................. 50 
4.2.2 Themes ................................................................................................................. 50 
4.2.3 Theoretical saturation ........................................................................................... 51 

4.3 FINDINGS ............................................................................................................................ 52 
4.3.1 Results on the potential antecedents of RPA adoption. ....................................... 52 
4.3.2 Technological factors ............................................................................................ 56 
4.3.3 Organisational factors ........................................................................................... 65 
4.3.4 Environmental factors ........................................................................................... 73 
4.3.5 Perceived ease of use (PEOU) and perceived usefulness (PU) .......................... 80 

4.4 SUMMARY OF CHAPTER 4 .................................................................................................... 81 
4.4.1 Technological factors ............................................................................................ 81 
4.4.2 Organisational factors ........................................................................................... 82 
4.4.3 Environmental factors ........................................................................................... 82 
4.4.4 Summary of findings ............................................................................................. 82 

CHAPTER 5. DISCUSSION OF THE FINDINGS ............................................. 84 
5.1 INTRODUCTION .................................................................................................................... 84 
5.2 INTERPRETATION OF FINDINGS ............................................................................................. 84 

5.2.1 Antecedents to RPA adoption .............................................................................. 84 
5.2.2 Technological factors ............................................................................................ 85 
5.2.3 Organisational factors ........................................................................................... 88 
5.2.4 Environmental factors ........................................................................................... 91 

5.3 SUMMARY OF CHAPTER 5 .................................................................................................... 93 
5.3.1 A revised conceptual model ................................................................................. 93 

CHAPTER 6. RECOMMENDATIONS AND CONCLUSION ............................ 95 
6.1 INTRODUCTION .................................................................................................................... 95 
6.2 LINKING TO THE RESEARCH QUESTIONS ................................................................................ 95 
6.3 CONTRIBUTIONS OF THE STUDY ............................................................................................ 97 

6.3.1 Industry ................................................................................................................. 97 
6.3.2 Academic .............................................................................................................. 97 

6.4 LIMITATIONS OF THE STUDY .................................................................................................. 98 
6.5 RECOMMENDATIONS FOR FUTURE RESEARCH ....................................................................... 98 
6.6 CONCLUSION ...................................................................................................................... 99 



ix 

REFERENCES ............................................................................................... 100 

APPENDIX A: Participant Information Sheet ............................................. 112 

APPENDIX B: Consent Form ....................................................................... 113 

APPENDIX C: Research Instrument ........................................................... 114 

APPENDIX D: Ethical Clearance ................................................................. 115 

APPENDIX E: NVIVO word cloud ................................................................ 116 

APPENDIX F: Thematic Analysis ................................................................ 117 

 



x 

LIST OF TABLES 

Table 2-1:Potential benefits of RPA (a review of literature) .............................. 13 

Table 2-2: Limitations of RPA (a review of literature) ....................................... 15 

Table 3-1: Participants of the study .................................................................. 40 

Table 3-2: Consistency table (interview questions and research questions) .... 42 

Table 4-1: Summary of initial codes per participant ......................................... 50 

Table 4-2: Summary of themes ........................................................................ 51 

Table 4-3: Summary of themes and codes ...................................................... 52 

Table 4-4: RPA implementation use cases in the South African insurance 

industry ............................................................................................................ 53 

Table 4-5: Potential antecedents to RPA adoption in the South African 

insurance industry ............................................................................................ 56 

Table 4-6: Results on how participants perceived relative advantage as a 

factor in RPA adoption. .................................................................................... 59 

Table 4-7: Results on how participants perceived the complexity of RPA as 

a factor in its adoption. ..................................................................................... 61 

Table 4-8: Results of how participants perceived compatibility as a factor in 

RPA adoption. .................................................................................................. 62 

Table 4-9: Results on how participants perceived the cost as a factor in 

RPA adoption. .................................................................................................. 65 

Table 4-10: Results on how participants perceived management support as 

a factor in RPA adoption. ................................................................................. 67 

Table 4-11: Results on how participants perceived the skills as a factor in 

RPA adoption. .................................................................................................. 69 



xi 

Table 4-12:  Results on how participants perceived the strategy as a factor 

in RPA adoption. .............................................................................................. 72 

Table 4-13:  Results on how participants perceived change management as 

a factor in the RPA adoption process. .............................................................. 73 

Table 4-14: Results on how participants perceived competition pressures as 

a factor influencing RPA adoption. ................................................................... 76 

Table 4-15: Results on how participants perceived government regulations as 

a factor influencing RPA adoption. ................................................................... 77 

Table 4-16: Results on how participants perceived vendor support as a factor 

in RPA adoption. .............................................................................................. 80 

Table 4-17: Summary of Findings .................................................................... 83 

 



xii 

LIST OF FIGURES 

Figure 2-1: Size of the South African insurance industry in the financial sector 10 

Figure 2-2: Technological, organisational, and environmental framework 

(Tornatzky & Fleischer, 1990) .......................................................................... 20 

Figure 2-3: Technology acceptance model (Davis et al., 1989; Venkatesh et 

al., 2003) .......................................................................................................... 23 

Figure 2-4: A general diffusion model (an adaptation of the diffusion of 

innovation theory by Rogers, 1995; Askarany, 2011) ....................................... 25 

Figure 2-5: TOE-DOI-TAM Conceptual Framework ......................................... 28 

Figure 3-1: Coding qualitative data - extract from Participant B's interview 

transcript .......................................................................................................... 44 

Figure 5-1: Revised TOE-DOI-TAM Conceptual Framework ........................... 94 

 

file:///C:/Users/27827/Desktop/THESES%202023/WITS/Assignments/Kunene,%20Namdipha%207%20Feb%2023/FINAL%209%20Feb%2023/Nkunene_%20Research%20Report%20_%20MMDB%202023%20(002),%20FINAL.docx%23_Toc126920565
file:///C:/Users/27827/Desktop/THESES%202023/WITS/Assignments/Kunene,%20Namdipha%207%20Feb%2023/FINAL%209%20Feb%2023/Nkunene_%20Research%20Report%20_%20MMDB%202023%20(002),%20FINAL.docx%23_Toc126920567


xiii 

LIST OF ACRONYMS 

AI   Artificial intelligence 

BPM   Business process management  

DOI   Diffusion of innovation theory  

NLP    Natural language processing 

PEOU   Perceived ease of use 

PU   Perceived usefulness 

RPA   Robotic process automation  

SA   South Africa 

TAM   Technology acceptance model  

TOE   Technology-organisation-environmental framework  

TOE-DOI-TAM  Conceptual framework 

 



1 

CHAPTER 1. INTRODUCTION 

1.1 Background of the study 

Automation can be defined as the application of technology to execute processes 

using programmed commands with minimal human input (Eling et al., 2022). 

Studies about the transformation of industries in this digital era have proposed 

automation and robotics as major enablers of technological transformation for 

most organisations (Ki et al., 2022).  

The concept of automation in the financial sector has been around for years; 

however, insurance companies have lagged in implementing automation 

technologies in their core processes (McKinsey Digital, 2017). The impact of 

Covid-19 has accelerated the need for the adoption of digital technologies and 

transformation of the global business market. The pandemic lockdown 

regulations and other mitigation measures that were implemented by the 

government affected how insurance organisations conduct business, which to a 

large extent traditionally required physical engagement and manual processes 

(Schlemmer et al., 2022). 

Robotic process automation (RPA) is an automation revolution that has disrupted 

every industry, and the insurance industry is no exception (Lacity et al., 2016) 

RPA is a software technology that uses rules and algorithms to mimic back-office 

tasks performed by human workers (Ghouse & Sipos, 2021). According to Katke 

et al. (2019), RPA affects the process automation landscape when compared to 

solutions like business process automation and workflows. RPA offers multiple 

benefits to insurers, from automating business processes; to reducing operational 

costs, seamless consumer experience, and the ability to remain competitive in 

this digital era (Schlemmer et al., 2022).  

 



2 

1.2 Research problem 

The South African insurance industry has been ranked as the least trusted 

industry within the financial sector (KPMG South Africa, 2021). With a negative 

0.4% average net sentiment score, insurance consumers cited the industry’s 

overall claims process as one of the major concerns – sentiments around the 

turnaround time of claims settlements and non-payment of claims have 

contributed to the dissatisfaction (PwC South Africa, 2021). According to Jones 

et al. (2019), the unsatisfactory customer services experienced by insurance 

customers can be attributed to insurers’ over-reliance on manual (traditional) 

processes, thus failing to meet customers’ needs. This data-intensive industry is 

overwhelmed with repetitive, time-consuming, and manual processes such as 

claims and underwriting which have negatively impacted consumer satisfaction 

and increased expenditure costs for daily operations (Kaur, 2022). 

From the literature it is understood that the insurance industry is perceived to be 

a laggard in the adoption of RPA. The industry was ranked the third-lowest 

industry to leverage RPA, heedless of the benefits derived from RPA adoption 

(Acceleration Economy, 2021). Lamberton et al. (2016) explored the impact of 

Robotics, RPA, and artificial intelligence (AI) on the insurance industry, and in 

their study, they identified common RPA adoption challenges faced by insurers 

such as choosing the correct processes to automate, strategic misalignment and 

incompatibility with legacy IT systems. Several studies have argued that for many 

insurance organisations the RPA adoption challenges extend beyond complex 

technology transitions, and much concern regarding data security and risks 

associated with RPA being implemented in their processes (Kaur, 2022; WNS, 

2018). Oza et al. (2020) associate the drawbacks of RPA adoption in the 

insurance industry with the fear of job replacement by automation, doubt around 

the precision and accuracy of tasks completed by bots, and high investment costs 

for implementing the technology. In the same vein, Smith (2020) found that slow 

adoption by insurers reflects the industry’s conservative culture towards 

disruptive technologies and changing business models.   

Though a considerable amount of global RPA implementation cases in insurance 

exists, many insurance organisations in South Africa are now looking at adopting 



3 

RPA to assist in automating their business processes. The heightened interest in 

RPA requires contributions to be made to the limited academic literature on RPA 

adoption within the South African insurance industry and creates a need to 

investigate the antecedents, and potential enabling or hindering factors 

influencing the adoption of RPA relevant to SA insurers.  

1.3 Statement of purpose  

The purpose of this study was to investigate the key antecedents and factors 

influencing the adoption of robotic process automation (RPA) within the South 

African insurance industry. Additionally, the study aimed to describe the influence 

of technological, organisational, and environmental factors from the perspective 

of the key decision-makers in the RPA adoption process.  

1.4 Research questions  

Drawing on the extant literature and the integration of the adoption models, the 

research questions were as follows:   

The main research questions:  

1. What are the potential antecedents to RPA adoption in the South African 

insurance industry? 

2. What factors influence the adoption of robotic process automation (RPA) 

in the South African insurance industry from a technology, organisational 

and environmental perspective?  

Research sub-questions:  

2.1  Which factors are perceived to have a positive influence on RPA adoption 

by the South African insurance industry? 

2.2  Which factors are perceived to have a negative influence on RPA 

adoption by the South African insurance industry?  

 



4 

1.5 Rationale  

Emerging technologies such as artificial technology and robotic process 

automation are changing the delivery of end-to-end insurance services. These 

technologies are providing insurers with enhanced capabilities to work smarter, 

faster, and more accurately (Cranfield & White, 2016). This research is aimed at 

understanding the potential antecedents and factors influencing robotic process 

automation adoption in the South African insurance industry.  

Studies about the implementation of RPA across various industries provide case 

examples of how these industries can acquire value from adopting automation 

technology (Ghouse & Sipos, 2021). According to a study conducted at Ernst & 

Young, the application of RPA introduces some challenges and opportunities for 

the insurance industry (Lamberton et al., 2016). The study points out that despite 

the cost reduction and efficiency benefits of RPA, several issues are encountered 

by insurers implementing RPA projects (Lamberton et al., 2016).  

The widespread RPA adoption makes the technology crucial to many sectors, 

thus requiring further research to be initiated (Oza et al., 2020). From a business 

perspective, it would be of interest to insurers to understand what factors to 

consider before the adoption to mitigate implementation challenges and increase 

technology uptake. The outcome of this research may also be used to identify 

common conditions driving the adoption, which may include industry trends and 

certain business processes that could imminently be automated by insurers to 

improve organisational efficiency and productivity.  

From a South African perspective, the data gathered from the participants will 

provide local insights on what factors SA insurers find to be either enabling or 

hindering RPA adoption.  

1.6 Delimitations of the study 

This study is based on robotic process automation (RPA) and excludes other 

types of automation technologies. Despite the widespread adoption of RPA 

across various industries within the financial services sector, this study focuses 



5 

especially on antecedents and factors influencing RPA adoption within the South 

African insurance industry. Due to time constraints, this study is unable to 

encompass all factors identified in the three theoretical frameworks chosen. The 

scope is restricted to factors closer to the study as identified in the conceptual 

framework. The study did not test the constructs of the conceptual framework; 

rather, it allowed the participants to share their perceptions on how identified 

factors influenced RPA adoption within the insurance industry. 

1.7 Definition of terms 

Adoption: refers to the acceptance and initial use of the technology (RPA) in the 

organisation.  

Adopters: refers to the insurance organisations (as listed in the Insurance 

Institute of South Africa).  

Digital native: a customer who has grown up in the presence of digital 

technology.  

Grey literature: refers to organisational literature produced outside academic 

and commercial publishing channels (Chugh et al., 2022).  

Innovation: refers to “an idea, practice, or object that is perceived as new by an 

individual/organisation” (Rogers, 1995). 

Image recognition: refers to the process of detecting an object from an image 

(Berruti et al., 2017). 

Insurers: defined as an organisation or enterprise that issues or sells insurance 

policies and provides financial cover to a customer in the form of a sum assured 

in case of occurrence of the event as stated in the insurance policy (Bharti AXA 

Life, n.d.).  

Intelligent automation: refers to the integration and use of multiple automation 

technologies such as robotic process automation with artificial intelligence to 

streamline business processes (Berruti et al., 2017). 



6 

Management: refers to a professional who occupies a leadership role and is 

responsible for a team of people within an insurance company.  

Robotic process automation (RPA): is defined as a software technology which 

uses software bots aimed at automating routine and repetitive human tasks 

(Ivančić et al., 2019). 

South African insurance industry: the industry involves companies and people 

who develop insurance policies, and sell, administer and regulate them. It 

consists of non-life insurers, life insurers and reinsurers (KPMG South Africa, 

2021).  

Technical expert: is limited to RPA vendors or companies selling RPA software 

technology to insurance companies.  

1.8 Assumptions 

The assumptions for this research are based on the reviewed literature pertaining 

to the adoption of RPA in the insurance industry. This research adopts an 

abductive approach, as it has drawn on past theoretical studies to develop the 

conceptual model and used empirical findings to modify the model. The following 

assumptions were made.  

An integration of adoption models (TOE, TAM and DOI) provides a 

comprehensive overview of factors influencing RPA adoption in the South African 

insurance industry (Julies & Tranos, 2021). In the South African insurance 

industry, customer-facing services such as claims registration and processing will 

be significantly disrupted by RPA adoption. Due to the competitive pressures 

within the industry, insurers have shifted to adopt digital technologies such as 

RPA to remain competitive. The research also assumed that the low RPA 

adoption rate can be attributed to the slow digital transformation of the industry 

and the lack of RPA-skilled experts in insurance organisations.  

 

 



7 

1.9 Chapter outline  

The overall structure of the study takes the form of six chapters as detailed below.  

Chapter 1: Introduction 

This chapter presents the background to the research on RPA adoption within 

the South African industry. The research problem and rationale for conducting 

the research are highlighted. It further identifies three research questions aimed 

at answering the research problem.  

Chapter 2: Literature review and theoretical framework  

This chapter provides a critical review of the available literature regarding robotic 

process automation and its adoption within the insurance industry. The chapter 

further examines the current body of knowledge on the technology adoption 

factors as identified in a technology-organisation-environmental framework 

(TOE), the technology acceptance model (TAM) and diffusion of innovation 

theory (DOI). The chapter concludes by proposing a conceptual framework based 

on the integration of the adoption models.  

Chapter 3: Research methodology  

This chapter provides an overview of the research design and methodology 

utilised in this research. The chapter also presents the data collection methods 

and ethical considerations.  

Chapter 4: Research findings and data analysis 

This chapter presents the empirical findings of this study and uses thematic 

analysis to analyse the empirical data gathered from the semi-structured 

interviews.  

Chapter 5: Discussion  

This chapter links the findings of the study with the research questions and then 

considers the implications of these on the conceptual framework.  



8 

Chapter 6: Conclusion and recommendations   

This chapter summarises the study and links the findings to the research 

questions. The chapter also includes the contributions of the study and suggests 

future research.  



9 

CHAPTER 2. LITERATURE REVIEW AND 

THEORETICAL FRAMEWORK 

2.1 Introduction 

This literature review offers an overview of the significant literature available 

regarding robotic process automation technology (RPA) and the adoption of RPA 

within the insurance industry. The literature review begins by defining automation 

and RPA. In describing the impact of RPA software technology in today’s 

business environment, this chapter provides examples of RPA benefits and 

limitations.  

The section of the literature review explores the role of RPA within the insurance 

industry and the antecedents of RPA adoption by insurers. Under the analytical 

framework, this study reviews the existing body of knowledge regarding 

technology adoption models, with a focus on technology-organisation-

environment (TOE), technology acceptance (TAM) and diffusion of innovations 

(DOI) and their relevance to this research. The literature review further explores 

the integration of adoption models to better understand the complexities of 

adopting technologies at an organisational level.  

Finally, the literature review concludes by providing a research model of factors 

influencing RPA adoption in SA’s insurance industry using theoretical 

frameworks. The critical literature review carried out used peer-reviewed journals 

and available research on the adoption of RPA. 

2.2 South African insurance industry  

The South African insurance industry is categorised as complex, concentrated, 

and competitive, with an active regional and global presence. As a vital pillar of 

the South African financial system, the insurance industry accounts for 18% of 

the financial sector and has the highest penetration in the African insurance 

market (International Monetary Fund, 2022).  



10 

SA’s insurance industry comprises 181 registered insurance companies which 

include life and non-life insurers, reinsurers, and captives (International Monetary 

Fund, 2022). Life insurers have a significant dominance in the market (approx. 

R 3 trillion GWP) whereas non-life insurance has approximately R 131.6 billion 

GWP. (KPMG, 2022).  

  

 

 

 

 

Figure 2-1: Size of the South African insurance industry in the financial sector 

Like the global insurance market, the South African insurance landscape is in a 

continual state of transition, faced with challenges such as a) the impact of the 

coronavirus pandemic on business processes; b) inflationary changes; c) 

decelerating prices owing to the emergence of Insurtech and shifting customer 

demands; d) disparate legacy systems; and e) demanding regulatory 

requirements (International Monetary Fund, 2022).  

2.3 Adoption of technology in the SA insurance industry 

The digitalisation of the South African insurance industry has become critical in 

recent years, with increased market pressures to launch digital-led services and 

the impact of Covid-19 on traditional business processes (Deloitte, 2021). Many 

insurance organisations are turning to robotic process automation to assist them 

with the digital transformation of their businesses (Pillay & Njenga, 2021). This 

view is supported by a survey conducted by KPMG that found that SA insurers 

who had embraced digital capabilities were able to operate seamlessly during 

COVID-19 and continue to serve changing consumer needs, whereas those 

insurers who have not explored digital capabilities are experiencing difficulties 

with remaining competitive in the current insurance market (KPMG South Africa, 



11 

2020). However, most insurance incumbents have demonstrated limited adoption 

of digital technologies, which has enabled Insurtech players such as Naked and 

Pineapple to gain traction in the market (Moodley, 2019). Insurtech offers 

competitive premium rates by reducing operational overheads through 

automation and adopting decentralised digital models and digital channels to 

improve customer experience (Moodley, 2019). 

2.4 Robotic process automation (RPA)  

The emergence of robotic process automation (RPA) technology can be traced 

back to the early 2000s; however, the term gained popularity in 2012 when 

industries displayed significant interest and uptake in the technology (Doguc, 

2020). In defining the term RPA, this study agrees that there is no single definition 

for robotic process automation (RPA) as the definition varies in the extant 

academic and professional literature. While much of the RPA research has 

proposed various definitions, for this research this study adopted the following 

definition: “RPA is a software technology used to automate manual and repetitive 

actions to complete single or multiple processes with minimal or no human 

intervention” (Armstrong & Lee, 2021; Ivančić et al., 2019).  

There seems to be a pervasive view that RPA can only be understood by 

describing its application in businesses. From a practical point of view, RPA 

combines several applications into one software technology that can be used to 

automate rule-based, repetitive, and manual office tasks (Chugh et al., 2022). In 

this study, Madakam et al. (2019) link the application of RPA to workflow 

automation in organisational processes, where the RPA software technology 

uses a graphical user interface to record and map out the actions (step by step) 

performed by a human, and then integrates those actions into programmed 

commands which enable the software bots to perform these actions 

automatically.  

However, there is incongruence in RPA studies as to what RPA can do, while 

many researchers claim that RPA only emulates the actions of a human by simply 

recording the steps taken to complete a task. However, this does not represent a 



12 

consensus view, as some studies argue that the capabilities of RPA extend 

beyond mimicking human actions, including manipulating and interpreting data 

and providing communication across multiple platforms (Hofman et al., 2020; 

Doguc, 2020). Given the varying capabilities of RPA, the next section will provide 

an overview of the different types of RPA software technology available for the 

organisation. Further, it will present benefits derived from using the technology 

and its limitations.  

2.5 Types of RPA, benefits, and limitations 

The progression of technology has led to the distinction of RPA software 

technologies that can be deployed by different organisations. The theoretical 

orientation of RPA is consistent with the industry orientation in distinguishing the 

types of RPA based on the capabilities of the RPA bot (Chugh et al., 2022; 

Ghouse & Sipos, 2021; Oza et al., 2020; WNS, 2018; Lamberton et al., 2016). 

An RPA bot can be seen as an intelligent software-based robot which uses 

programmed algorithms to perform different tasks (Chugh et al., 2022), and not 

a physical robot.  

2.5.1 Types of RPA  

The literature has identified the types of RPA that are commonly used as 

attended, unattended and hybrid:  

a) Attended RPA 

Attended RPA requires human intervention to complete the intended automation. 

The RPA bot will perform tasks triggered by the human user or a system-level 

event and the automation takes place at an individual workstation (Choi et al., 

2021; Ghouse & Sipos, 2021). 

b) Unattended RPA  

In an unattended RPA, the RPA bots are designed to execute back-office tasks 

and interact with the technology with very little and – in most cases – no human 

intervention. This automation can be scheduled, operates on the enterprise’s 



13 

server, and is suitable for rule-based tasks which do not require human input 

(Choi et al., 2021; Ghouse & Sipos, 2021). 

However, a recent study and a report by a widely used RPA vendor (Chugh et 

al., 2022; UiPath, 2020a) extend the literature regarding RPA types by introducing 

hybrid RPA bots. They describe hybrid RPA as a combination of attended and 

unattended bots used to provide a single end-to-end automation solution for 

complex business tasks. This type of RPA can operate at both desktop and server 

levels, thus providing flexibility and scalability to an organisation (Chugh et al., 

2022; UiPath, 2020a). As the literature positions the types of RPA based on the 

level of human interactions with the technology, this study suggests that for 

organisations intending to adopt RPA, their distinction between types of RPA 

should adhere to the sophistication level of the RPA.  

2.5.2 RPA benefits  

The question of how adopting and using RPA can benefit an organisation is 

widely researched. The table below draws extant examples from reviewed 

literature and provides multiple benefits derived from RPA adoption and usage 

within an organisation.  

Table 2-1:Potential benefits of RPA (a review of literature) 

Potential RPA Benefits  Authors 

Cost reduction 
(Chugh et al., 2022; Armstrong & Lee, 2021; Choi et al., 2021; 
Ghouse & Sipos, 2021; Pekonen & Lähteinen, 2021; Doguc, 
2020; Gami et al., 2019) 

Improved process 
efficiencies 

(Chugh et al., 2022; Armstrong & Lee, 2021; Choi et al., 2021; 
Pekonen & Lähteinen, 2021; Doguc, 2020; Gami et al., 2019) 

User-friendliness 
(Chugh et al., 2022; Armstrong & Lee, 2021; Ghouse & Sipos, 
2021; Gami et al., 2019) 

Improves quality (reduced 

errors) 

(Chugh et al., 2022; Armstrong & Lee, 2021; Choi et al., 2021; 

Ghouse & Sipos, 2021; Pekonen & Lähteinen, 2021; Doguc, 

2020; Gami et al., 2019) 



14 

Flexibility and scalability (Chugh et al., 2022; Ghouse & Sipos, 2021; Doguc, 2020) 

Increases customer 

satisfaction 

(Chugh et al., 2022; Armstrong & Lee, 2021; Pekonen & 

Lähteinen, 2021; Doguc, 2020; Gami et al., 2019) 

Improved compliance and 

governance 
(Chugh et al., 2022; Armstrong & Lee, 2021; Gami et al., 2019) 

Compatibility with 

existing IT architecture 

(Chugh et al., 2022; Choi et al., 2021; Pekonen & Lähteinen, 

2021; Doguc, 2020; Gami et al., 2019) 

 

As Gami et al. (2019) suggest, to organisations, RPA is a simple automation 

solution aimed at resolving operational issues. In resolving these operational 

issues, the literature offers a wide range of benefits to organisations as cited in 

the above table. In the insurance industry, RPA has been deployed in various 

processes and has brought many benefits to insurers. The most significant 

benefits of RPA in insurance are the reduction in operating costs and 

improvement in operational performance as noted by Sreeja and Shanthini 

(2018). In the same vein, this study suggests that insurance organisations with 

the business goal of reducing costs, improving operational efficiencies, and 

increasing service quality and customer satisfaction can largely benefit from 

adopting RPA.  

2.5.3 Limitations of RPA 

Despite many RPA benefits presented, authors have argued that RPA is not 

without limitations. The limitations of RPA have been identified by determining 

what RPA is not in comparison with other automation solutions. These 

comparisons of RPA with AI and business process management (BPM) indicate 

the limitations of the technology (Armstrong & Lee, 2021; Doguc, 2020) as 

displayed in the table below.  

 



15 

Table 2-2: Limitations of RPA (a review of literature)  

Limitations  Authors  

Dependency on standardised 

and rule-based tasks 

(Armstrong & Lee, 2021; Choi et al., 2021; Fernandez & 

Aman, 2021; Hofmann et al., 2020; Vijaya et al., 2019) 

Requires intelligent automation 

(e.g., embedded with natural 

language processing (NLP) and 

image recognition to process 

unstructured data 

(Pekonen & Lähteinen, 2021; Hofmann et al., 2020) 

Requires human intervention to 

control and monitor RPA bots 

(Choi et al., 2021; Fernandez & Aman, 2021; Hofmann et 

al., 2020) 

Requires frequent maintenance 
(Armstrong & Lee, 2021; Fernandez & Aman, 2021; 

Hofmann et al., 2020) 

 

Although RPA is widely implemented in conjunction with other digital 

technologies such as AI and BPM, adopting the technology alone poses some 

limitations to the organisation. Choi et al. (2021) suggest that organisations 

intending to adopt the technology need to outline what is included in the scope of 

RPA in conjunction with what the technology can do. Despite the limitations 

presented in the literature; strategy is identified as a tool to overcome these 

limitations of RPA. This strategy will address the scope of the RPA 

implementation (short-term or long-term), which technologies can be 

implemented to supplement the drawbacks of RPA, what effect RPA will have on 

business processes and the required alignment of business, and RPA strategy 

within an organisation. 

2.6 RPA in insurance   

Insurers are recognising robotic process automation as an innovative force 

disrupting their traditional work environment. In a CEO survey conducted by PWC 



16 

with various insurance organisations, it was acknowledged that RPA has become 

critical across many insurance processes (WNS, 2018). The implementation of 

RPA in various insurance processes has been researched globally; however, 

limited studies have been conducted in South Africa, as seen in the review 

literature. For instance, studies explored RPA implementation in insurance claims 

registration and processing and identified that adopting RPA leads to faster 

claims processing, increased data accuracy and enhanced customer experience 

(Oza et al., 2020; Vijaya et al., 2019). The adoption of RPA can help streamline 

processes such as underwriting, claims and fraud detection, and customer-facing 

services (WNS, 2018).  

While much of the RPA research in insurance has highlighted various use cases, 

most studies have been limited to providing a process perspective on the 

adoption as they focused on siloed processes. A report by Capgemini (2017) 

provided a holistic overview of the advantages of RPA adoption in the insurance 

industry. They categorised these advantages into three perspectives; a) 

productivity perspective – reduces production downtime as the application can 

run 24/7; b) operational perspective – reduces cycle-time of the process by 75%, 

reduces output errors, ensures higher security and continuity of processes; c) 

regulatory perspective – ensures regulatory compliance when implemented in 

certain administrative and legal processes. There appears to be an important 

need for insurers intending on adopting RPA to understand how the adoption of 

technology will impact the overall business. 

2.6.1 Antecedents to RPA adoption  

Most of the literature that focused on the adoption of technology at an 

organisational level has suggested a few antecedents influencing the adoption. 

Saghafian et al. (2021) argue that there is an inconsistency pertaining to the 

antecedents that influence technology adoption. They can be viewed as 

conditions contributing to the success or failure of the adoption. For this study, 

antecedents will relate to preceding conditions considered to be influencing RPA 

adoption in the insurance industry.   

 



17 

a) The high volume of manual repetitive process  

The insurance industry is known as a data-intensive industry with several manual 

processes. This makes RPA adoption ideal in insurance – for automating manual 

repetitive tasks prone to human errors, rule-based and high data volume tasks 

(Sobczak, 2021). There are numerous benefits that insurers can derive from 

adopting RPA in their organisations, such as improved service quality, cost-

saving, and scalability, as cited in section 2.5.2 RPA benefits. However, selecting 

the optimal processes ideal for automation can be seen as a challenge for most 

insurers (Lamberton et al., 2016).  

b) Regulatory compliance  

The insurance industry is a highly regulated environment that requires high levels 

of compliance where the slightest oversight may cause financial loss to an 

insurance organisation. For insurers, RPA ensures that regulatory compliance is 

met through automated screening, early detection of missing documents, 

expediting verification of know your customer (KYC) compliance, and detecting 

fraudulent data. Considering the threats of cybercrime, the implementation of a 

cloud-native RPA solution may pose major data privacy and security for insurers 

as sensitive information may be exposed (Katke et al., 2019).  

c) Competition intensity 

The digital era has changed the way insurers conduct business and with the rise 

of Insurtech, it has become apparent that incumbents need to adopt digital 

technologies to remain competitive. The effects of COVID-19 and competition 

intensity are pressuring insurance organisations to explore and adopt RPA to 

assist them in digitally transforming their processes. According to Lamberton et 

al. (2016), RPA offers higher efficiency and lower operating costs, which allows 

insurance organisations to transform their processes and build a competitive 

advantage.  

 

 



18 

d) Seamless integration  

Lamberton et al. (2016) describe RPA as a non-invasive technology that allows 

for seamless integration with existing IT systems within an organisation and faster 

implementation. The literature has highlighted that many insurance organisations 

still use old legacy systems, and changing the entire IT infrastructure to 

accommodate a new technology seems like a huge investment that not many are 

willing to make. RPA vendors have argued that RPA is a non-invasive technology 

that is compatible with existing IT infrastructure (UiPath, 2020a). Based on RPA’s 

non-invasive compatibility, it is more attractive to insurers that want a technology 

adoption process with minimal disruptions to normal business operations.  

e) Organisational culture 

Resistance to change from employees and a lack of management support can 

result in RPA adoption failures and low usage of the technology within an 

organisation (Sobczak, 2021). There is a strong chance that the organisational 

culture may feel threatened by the adoption of new technology, which may pose 

changes to the existing culture. This view is supported by Lamberton et al. (2016), 

who write that the success of RPA adoption within an organisation requires a 

supportive organisational culture that embraces change. Likewise, this study 

agrees that insurers must utilise change agents to assist in the journey of 

transitioning the existing culture to be supportive of RPA adoption.  

f) Creating value for customers 

Insurers are faced with high expectations from digital natives, who expect faster 

services, paperless processes, reduced costs, and digital engagements. Many 

insurance organisations cannot afford to lose any more customers due to 

paperwork (such as KYC, policy documents, and claims) or disengaged manual 

processes, and are seeking solutions. RPA promises to deliver automated, 

simplified onboarding processes, faster response rates, and cost benefits through 

automation. However, a study by Cenfri and Equisoft that explored the effects of 

automation within the African insurance industry acknowledged the presence of 

a digital divide amongst African customers and proposed that insurers seeking to 

adopt RPA should balance automated services with physical engagements to 



19 

create value for both non-digital savvy and traditional customers (Schlemmer et 

al., 2022). 

However, the literature is not clear as to which of these antecedents lean towards 

creating a case for RPA adoption in insurance. It is noted that some of the 

identified antecedents may not apply to the South African insurance landscape, 

as most of the reviewed literature is not country specific. In bridging the gap 

identified, this study aims to contribute to the literature by identifying the 

antecedents relevant to the South African insurance landscape.  

P1: High volume of manual processes, demands from customers and competition 

intensity are the key antecedents to RPA adoption in the insurance industry.  

2.7 ANALYTICAL FRAMEWORK 

Studies show that the impact of technology adoption on organisations is 

unpredictable. Some authors have suggested that a newly adopted technology 

creates opportunities for organisations (Gangwar et al., 2014), whereas others 

argue that a newly adopted technology may disrupt existing processes, thus 

having a negative impact on the organisation (Dube et al., 2020). 

Dube et al. (2020) define technology adoption as the acceptance and use of new 

technology by an individual, group or organisation. Studies conducted in the past 

few decades reveal that technology adoption is an extensively studied area in 

information systems with several technology adoption models introduced by 

numerous researchers (Fonseka et al., 2020).  

This research focuses on the organisation-level perspective of technology 

adoption, thus the selection of reviewing the technology-organisation-

environment framework (TOE), technology acceptance model (TAM), and 

diffusion of innovations theory (DOI). 

2.7.1 Technology-organisation-environmental framework (TOE) 

The technology-organisation-environment framework (TOE) is aimed at providing 

a theoretical lens for the factors influencing IT adoption in organisations 



20 

(Tornatzky & Fleischer, 1990; cited by Rosario Oliveira Martins et al., 2011). TOE 

is regarded as the most utilised framework by researchers conducting studies on 

technology adoption at the organisational level (Fonseka et al., 2020). Several 

studies have revealed that the TOE framework is utilised to gain a better 

understanding of various IT adoptions such as e-commerce, enterprise resource 

planning (ERP), electronic data interchange (EDI), and knowledge management 

systems (Gangwar et al., 2014).  

 

Figure 2-2: Technological, organisational, and environmental framework (Tornatzky & 

Fleischer, 1990) 

The framework identifies three types of contexts – technological, organisational, 

and environmental – that may influence the adoption and implementation process 

of new technology by an organisation.   

a) Technological context  

The technological context considers both internal and external technologies 

available to the organisation (Fonseka et al., 2020). In this context, the 

characteristics, competence, and perceived benefits of the technology are cited 

as significant factors in RPA adoption studies (Gangwar et al., 2015; Julies & 

Tranos, 2021).  



21 

b) Organisational Context 

The organisational context relates to the organisation’s size, scope, and 

managerial structure and beliefs (Fonseka et al., 2020; Salwani et al., 2009). 

Studies have also identified managerial support, degree of centralisation, 

standardisation of processes, quality of talent, strategic use of technology, trust 

and organisational slack as significant factors in this context that inhibit and 

contribute to the adoption of innovations (Iraq, 2020; Julies & Tranos, 2021). 

However, the factors introduced by Tornatzky and Fleischer fail to include the 

financial resources available to the organisation, which may influence the 

organisational inclination to invest and adopt new technologies.  

c) Environmental context 

This context focuses on the external environment in which the organisation 

conducts its operations (Salwani et al., 2009). It identifies factors external to the 

organisation such as industry competition, suppliers' support, and governmental 

laws as significant factors influencing organisations to adopt technological 

innovations (Dube et al., 2020; Fonseka et al., 2020; Salwani et al., 2009).  

d) Limitations of technology-organisation-environmental framework 

(TOE)  

Despite being cited as the most utilised framework for organisational adoption, 

there are limitations associated with TOE. For instance, the list of factors 

identified in TOE is not succinct and their application varies between the three 

contexts of the framework (Julies & Tranos, 2021). The framework does not 

provide information regarding the relationship between the contexts (Awa et al., 

2017). In addition, it appears that TOE places less emphasis role of the decision-

makers as a factor in the organisational context. Based on the RPA it is apparent 

that management has a significant influence on the adoption process of RPA.  

For this study, the TOE framework provides value-added information relevant to 

explaining and a theoretical foundation of the factors influencing RPA adoption 

by insurance organisations. However, the integration of the TOE framework with 

other technology adoption models is supported.  



22 

2.7.2 Technology acceptance model (TAM)  

The technology acceptance model (TAM) was developed by Davis (1989) based 

on the theory of reasoned action. TAM was commonly used to explain the user’s 

behaviour, beliefs, and attitudes toward accepting and adopting new 

technological innovations (Pires et al., 2011). The original constructs of TAM 

suggested five concepts: perceived usefulness (PU) and perceived ease of use 

(PEOU) that are linked to the users’ attitude (ATU) towards using new technology, 

the behavioural intention (BI) to use, and ultimately the actual use (AU) of new 

technology (Fonseka et al., 2020). Perceived usefulness (PU) is defined as a 

relative advantage based on the probability that an employee of an organisation 

will adopt a technology if it is confirmed that using the technology will enhance 

their work performance (Pires et al., 2011). Perceived ease of use (PEOU) relates 

to employees’ perception that the use of a certain technology will require less 

mental and physical effort (Pires et al., 2011).   

Most research has focused on the extension of TAM in efforts to expand the 

constructs of the model by introducing new external variables that influence the 

user to accept a new technology ((Hoong et al., 2017; Kauffman & 

Techatassanasoontorn, 2011; R. Sharma & Mishra, 2014). The extensions of 

TAM, such as TAM 2 and TAM 3 (see Figure 2-3) emphasised external variables 

(individual and social factors) that influence the adoption of ICT at an individual- 

level and considered the removal of attitude as a construct to increase the effect 

of PU and PEOU on behavioural intention (Hoong et al., 2017).  

However, from an organisational adoption of technology context, Dube et al. 

(2020) found the original TAM an appropriate model for predicting the adoption 

and usage of new technology at an organisational level. In their study, a causal 

relationship was suggested between the two constructs (PU and PEOU) of the 

model. It was also suggested that perceived ease of use of technology influences 

perceived usefulness, which implies that the simplicity of technology makes it 

more useful to the user.  



23 

 

Figure 2-3: Technology acceptance model (Davis et al., 1989; Venkatesh et al., 2003) 

a) Limitations of TAM  

Extensions of TAM were not conceptualised to address the adoption of 

technology in an organisational context but rather the individual’s acceptance of 

new technology (Julies & Tranos, 2021). This relates to the commonly cited 

limitation of TAM in the literature being self-report usage; in that way the model 

illustrates an individualistic view for PEOU and PU of a new technology which is 

independent of the usage of others (Koul & Eydgahi, 2017).  

Although not focused on individual intent to adopt new technology, this study finds 

it important for the participants to share their experiences and how they perceive 

the new technology as this influences the intention to adopt and ultimately use 

the technology. Drawing on Dube et al.’s (2020) concept of employing TAM in 

understanding organisational level adoption, this study considers the constructs 

of PU and PEOU as constructs that describe the organisational perceptions 

towards the acceptance of RPA technology, thus affecting the decision to adopt 

RPA.   

 

 



24 

2.7.3 Diffusion of innovations theory (DOI)  

Rogers (1995) introduced the diffusion of innovations theory (DOI) that identified 

innovation, communication channels and the social system as determinants of 

adoption. The theory defines diffusion as a process that involves an innovation 

being communicated over time amongst the members of a social system using 

specific channels (ibid). In an organisational context, innovation can be described 

as an idea, behaviour, or process new to the organisation (Damanpour & 

Gopalakrishnan, 1998). The authors further suggest two streams of diffusion of 

innovation in an organisation: a) generation of innovation – which relates to the 

internal generation of innovations to be solely used by the organisation; and b) 

adoption of innovation – which relates to externally developed innovations 

imported by an organisation (Damanpour & Gopalakrishnan, 1998). 

Over the years, organisations have adopted the latest techniques, practices, and 

technologies to remain competitive in the fast-paced business environment. This 

has placed greater emphasis on organisations to understand the diffusion of 

innovations to cope with the changes (Askarany, 2011). In addressing diffusion, 

Rogers (1995) describes the five characteristics of innovations: relative 

advantage, compatibility, complexity, trialability, and observability. These 

characteristics introduce a variety of factors influencing the adoption of 

innovations in an organisation, such as the level of uncertainty regarding the 

innovation; cost of the innovation; the expected return of investment associated 

with the innovation; flexibility; perceived benefits; availability of an innovation; 

knowledge about the innovation; and lastly, the type of innovation influences the 

adoption (Askarany, 2003).  



25 

Figure 2-4: A general diffusion model (an adaptation of the diffusion of innovation theory 

by Rogers, 1995; Askarany, 2011) 

a) Limitations of the diffusion of limitations theory (DOI) 

While DOI presents a case for the antecedents of technology, it does not 

distinguish between the causes for the non-adoption or adoption of a technology 

(MacVaugh & Schiavone, 2010). According to Oliveira & Martins (2011), DOI 

excludes the analyses of the environmental context in its theory, and the 

exclusions of environmental variables such as competition pressures limit the 

theory in explaining intra-organisation innovation adoption.  

Therefore, in this study, Askarany's (2011) general diffusion model (which is an 

adaptation of DOI) is seen as a more relevant model for understanding the 

influencing factors of RPA adoption by an insurance organisation. It is viewed as 

providing a comprehensive understanding of the technology adoption process 

and factors influencing the adoption. However, the focus will be on the attributes 

of innovation and adopters. The attributes of innovation will be limited to 

complexity, compatibility, and organisational strategy, and will only be considered 

as attributes of adopters.  



26 

2.7.4 Integration of the adoption models  

Various empirical and conceptual studies have shown the relevance of using 

TOE, TAM and DOI in understanding the factors affecting technology adoption. 

Each model has presented limitations that have justified the integration of models 

(Awa et al., 2012; Gangwar et al., 2015; Iraq, 2020). Iraq (2020) posited that the 

integration of various adoption models could increase the adoption of innovative 

technologies in organisations. 

Julies and Tranos (2021) claim that when compared, the constructs of the TOE 

and DOE show consistencies in models, for instance; a) DOI’s attributes of 

innovations with TOE’s technological context; and b) DOI’s attributes of adopters 

with the organisational context of the TOE. TAM and TOE are regarded as 

complementary models that have been integrated by previous studies. A TAM-

TOE model was used in determining the factors of cloud computing adoption in 

organisations (Gangwar et al., 2015). Another study used a TAM-TOE model to 

explain e-commerce adoption by SMEs (Awa et al., 2012).  

Several studies have suggested the integration of multiple technology adoption 

models and theories. In a South African context, a study that reviewed the 

progression of TAM and TOE suggested an approach to integrate the two models 

(Julies & Tranos, 2021). Another study considered the constructs of TAM, TOE 

and DOI to better understand the adoption of job automation technologies across 

various industries (Abdulla, 2019). In support of integrating multiple adoption 

models, Nesindande, (2020) incorporated TAM, DOI and motivational model into 

one framework to understand the experiences of human resources concerning 

the implementation of RPA in commercial banking (Nesindande, 2020). In the 

same vein, this study integrates the three models into a conceptual framework 

aimed at obtaining a comprehensive understanding of the factors influencing 

RPA adoption in SA’s insurance industry, as discussed in the following section.  

 



27 

2.8 Conceptual framework 

Based on the review of multiple theories and models regarding the adoption of 

new technology at an organisational level, a framework for identifying factors 

influencing the decision to adopt RPA was developed (see Figure 2-5). The 

framework acknowledges that the ultimate adoption of new technology within an 

organisation requires a phased approach. Drawing from the existing literature on 

the adoption process, this study adopted a four-staged adoption process, an 

adaptation of the five stages in the innovation-decision process by Rogers (2003). 

The framework acknowledged that different factors influence the adoption of new 

technology at different stages of the adoption process. The first stage of the 

adoption process (referred to as Knowledge) adopted the external and internal 

variables selected from DOI’s innovation characteristics and TOE framework and 

were framed to complement the RPA adoption process. The second stage 

(referred to as Piloting) considered organisational behaviour towards the 

technology after a trial implementation (initial human-technology interaction). 

Then, stage three of the framework considered perceptions of these factors on 

the decision to adopt or reject RPA within an insurance organisation. Lastly, 

dependent on the decision to adopt, the final stage of the framework considers 

the adoption and use of RPA in the insurance organisation.   



28 

 

 

Figure 2-5: TOE-DOI-TAM Conceptual Framework 

Solid lines (shapes) – represent independent and dependent variables.  
Dashed lines (shapes) – represent mediating variables.  



29 

2.8.1 Adoption process  

The adoption process is an adaptation of DOI’s innovation-decision process 

model which highlights the stages that an adopter goes through when adopting 

new technology. For this study, the RPA adoption process followed four stages.  

a) Knowledge: This stage involves two steps: a) a knowledge gathering 

step, which is heavily influenced by external factors that spark interest and 

awareness of technology for the potential adopter; b) the evaluation step, 

which involves that the adopter will be using information obtained to create 

a business case for the adoption of the technology.  

 

b) Piloting (trial implementation): this stage refers to the small-scale 

implementation, the initial organisation-technology interaction where the 

viability of the technology is tested. The success of this phase is heavily 

influenced by whether the technology is perceived to be easy to use and 

useful to the organisation (attitude formation).   

 

c) Decision: this stage leads to either the adoption or non-adoption of the 

technology. 

 

d) Adoption:  at this stage, the adopter acts upon the decision taken. In the 

case of deciding to adopt RPA, this phase will include the acceptance and 

use of RPA by the insurance organisation.  

In understanding the contextual factors influencing the adoption of RPA by SA 

insurance organisations, the following factors have been identified.  

2.8.2 Technological factors  

In terms of technological factors, perceived benefits, complexity, and 

compatibility of RPA are considered for this study. The relative advantage in this 

study refers to the perceived benefits of RPA. These benefits, derived from RPA 

implementation in insurance, extend beyond improving efficiency and cost 

reduction (Capgemini, 2017; Lamberton et al., 2016). The expectation from RPA 



30 

is that through the adoption, employees will be relieved of repetitive and manual 

tasks to focus on higher-value tasks; however, recent RPA cases in insurance 

show benefits such as personalisation and selling opportunities, cost advantage 

(savings) over RPA laggards, and enhanced customer experience through self-

service processes (WNS, 2018; Oza et al., 2020). This study then proposes:  

P2a: Relative advantage is positively associated with the perceived usefulness 
(PU) and perceived ease of use (PEOU) of RPA. 

Complexity relates to the level of difficulty in using and understanding a new 

technology by the potential adopters (Askarany, 2011). For this study, complexity 

was associated with the user-friendliness of RPA and easy-to-follow automation 

processes. Lanfranchi and Grassi (2022) warn against undocumented RPA 

processes which contribute to the complexity experienced by users of the 

technology. In their study, they found that when compared with other automation 

technologies like artificial intelligence and machine learning, RPA is regarded as 

a simpler automation technology; hence, the following proposition is made:  

P2b: Complexity is positively associated with the perceived ease of use (PEOU) 
of RPA. 

Compatibility refers to the level of consistency of innovative technology with 

legacy systems (Askarany, 2011). In the RPA adoption, this may relate to how 

compatible RPA is with the IT /internal systems of the organisation and whether 

the adoption will cause major interruptions to normal business operations. The 

perception that RPA is a non-invasive technology makes it compatible with many 

IT existing systems (Ivančić et al., 2019). However, a study by KPMG highlights 

that, for many South African insurance organisations, there is a high usage of old 

legacy systems that may not be compatible with newer technologies such as 

RPA. Furthermore, the study suggests that insurers may be required to make 

investments towards updating or changing the current IT infrastructure (KPMG 

South Africa, 2021). Based on the literature reviewed, the following proposition is 

made: 

P2c: Compatibility is negatively associated with the perceived usefulness (PU) 
and perceived ease of use (PEOU) of RPA. 



31 

2.8.3 Organisational factors 

Considering the skills required to effectively use RPA, it is the technology’s 

simple nature and the fact that little/no formal programming skills are required to 

use RPA that make it attractive to many organisations intending on automating 

their processes (WNS, 2018). However, some studies argue that the initial setup 

of RPA and automating more complex processes will require expert skills and 

capabilities (Lacity et al., 2016; Lamberton et al., 2016). While RPA research 

debates the skills required for RPA, a recent report on the future of automation in 

South Africa highlighted two skills issues: RPA skills shortages in South Africa, 

and a high migration rate of RPA professionals leaving South Africa for 

international opportunities(Burrows, 2022). This is supported by earlier 

observations that suggested that there is a short supply of RPA developers in 

South Africa (Solutioneers, 2020). Based on the literature, this study assumes 

that a shortage of skills in the current SA insurance industry will negatively affect 

RPA adoption, thus the following proposition is made:   

P3a: The current skills in the South African insurance industry are negatively 
associated with the perceived ease of use (PEOU) of RPA. 

In a previous study, management support was regarded as a significant factor 

in the decision to adopt or non-adopt RPA within the organisation (Juntunen, 

2018). Managerial support acts as a facilitating factor in the adoption process 

(acceptance and use) of RPA in the insurance industry, but it is worth noting the 

failure and non-adoption of RPA are high when there is no management buy-in 

from the start (Lamberton et al., 2016). Consequently, management support is 

considered a significant factor influencing the perceived usefulness of RPA within 

the insurance industry, thus the following proposition is made:  

P3b: Management support is positively associated with the perceived usefulness 
(PU) of RPA. 

 

The Covid-19 pandemic has accelerated that digital shift amongst SA’s insurance 

organisations, with most SA insurers adopting clear and aggressive 

modernisation strategies, from the digitalisation of processes (paving the way for 

RPA and AI usage) to creating customer-centric products (Kean, 2020).  



32 

Strategy refers to an organisational strategic approach to selecting business 

processes to automate and RPA-business alignment (Lamberton et al., 2016). In 

their study, Lamberton et al. (2016) listed strategy as one of the challenges in the 

RPA implementation processes, citing that many insurance organisations lack 

the business-IT strategic intent, and hence target wrong or fewer value processes 

to automate, which results in higher RPA implementation costs and requires 

additional efforts from employees. In support, Choi et al. (2021) state that 

organisations vaguely define the RPA adoption plans, citing the complexities in 

strategic planning. However, the study by Choi et al. (2021) highlighted that the 

absence of an RPA strategy may increase failure in RPA projects. Rawashdeh et 

al. (2022) argue for the importance of an RPA strategy, and further suggest that 

organisations intending on adopting RPA should begin with a strategy to realise 

the benefits of the technology. However, based on the literature reviewed, 

strategy is perceived as a challenge for most organisations adopting RPA, hence 

the following proposition is made: 

P3c: Strategy is negatively associated with the perceived usefulness (PU) of 
RPA. 

2.8.4 Environmental factors  

Competition pressures refer to pressures from other organisations that have 

adopted RPA within the insurance industry. The role of Insurtech and micro-

insurance is recognised as a positive motivator that places pressure on insurance 

incumbents to adopt digital technologies (Moodley, 2019).  

The South African Insurance Outlook 2021, a study conducted by Deloitte, found 

that within the insurance industry, first movers to adopt digital technologies in 

their organisations tend to maintain a competitive advantage over the digital 

laggards. These early adopters were also found to have seamlessly operated 

during the pandemic (Deloitte, 2021). This notion is also applicable in the context 

of insurers adopting automation technologies such as RPA. Rawashdeh et al. 

(2022) say that while there are early adopters in every industry, most companies 

have evident mimetic isomorphism in their adoption of RPA. Based on the 

literature reviewed, the following proposition is made: 



33 

P4a: Competition pressures within the SA insurance industry are positively 
associated with the perceived usefulness (PU) of RPA. 

 

Vendor support refers to the service providers of RPA that assist insurance 

organisations with the adoption and successful implementation of RPA projects 

(OpenSky Data Systems, 2020). RPA vendor selection is identified as a 

challenge for most financial and insurance industries implementing RPA. Many 

seek business partners that will provide a reliable product at a reasonable price 

and provide post-sale support (Fox et al., 2021). For many insurers seeking to 

adopt RPA, vendor support is a key demand in resolving data protection and 

other security concerns relating to RPA. The support extends to vendors 

providing adequate post-sale support and training to ensure effective utilisation 

of the RPA software technology. Based on the literature reviewed, the following 

proposition is made: 

P4b: Vendor support is positively associated with the perceived usefulness (PU) 
and perceived ease of use (PEOU) of RPA. 

 

In the South African context, government regulations lag as opposed to the fast-

changing technology landscape (Farhat, 2019). Government regulations relate 

to laws organisations need to consider before adopting RPA. Katke et al. (2019) 

highlight the complexity of automation within the financial services sector and the 

impact it may have on customers. In their study, it is suggested that adopters of 

automation need to study and comply with regulatory laws to safeguard the 

interests of customers. However, some studies acknowledge that statutory 

obligations may impact adoption but have found this factor as a less significant 

influence on RPA adoption (Tew, 2019; Katke et al., 2019). Concerns about data 

privacy and the legal implications of using RPA may negatively affect the 

adoption; hence, the following proposition is made:  

P4c: South African government regulations are negatively associated with the 

perceived usefulness (PU) of RPA. 



34 

2.8.5 Perceived ease of use (PEOU) and perceived usefulness (PU) 

of RPA  

PEOU describes the users’ perception that new technology is easy to use and 

user-friendly (Gangwar et al., 2015). This study considers the following 

fundamentals in understanding how the users of RPA perceive the technology to 

be easy to use or not: a) there are factors identified as influencing the perceived 

ease of use; b) the study assumes RPA’s ease of use can only be evaluated on 

the user’s initial interaction with the technology, which is during the piloting stage. 

This conceptual framework agrees with the view that a technology regarded as 

easy to use and easy to understand is perceived to be more useful to the adopter; 

hence it is likely to be adopted. For this study, PEOU and PU are regarded as 

mediating variables. This entails that the influence of TOE factors on the decision 

to adopt will be significantly influenced by how RPA software technology is 

perceived by the user.  

2.9 Summary of Chapter 2 

Existing literature has demonstrated the importance of insurance organisations 

making investments towards digital technologies like RPA to remain competitive. 

In summary, this chapter has presented the literature around RPA, the types of 

RPA applications, and the benefits and limitations of RPA. It has also highlighted 

that, although RPA has been implemented within the insurance industry, there 

are limitations in academic research on RPA in a South African context. 

Furthermore, a critical review of the TOE framework, technology acceptance 

model and diffusion of innovations theory has been presented to provide a 

theoretical orientation to this study. This study found that the application of the 

models individually resulted in limitations, hence the proposal of an integrated 

conceptual framework. The conceptual framework presented aims to contribute 

to the academic literature on RPA adoption whilst giving local insights on factors 

affecting the adoption of RPA from a SA insurers’ perspective. To date, several 

studies have suggested a multi-perspective model be used for a comprehensive 

view of the factors influencing technology adoption at an organisational level.  



35 

CHAPTER 3. RESEARCH METHODOLOGY 

This chapter describes the methodological approach adopted in this study. The 

chapter is divided into sub-sections that include the selected research design and 

methodology, followed by the data collection method. Thereafter, a discussion on 

the chosen data analysis strategy and interpretation of findings is presented. This 

chapter will also present the limitations of the chosen methodology and provide 

details on the quality assurance of this study and, finally, the ethical 

considerations.  

3.1 Research methodology and design  

Creswell (2007) describes the philosophical assumptions of the research as the 

researcher’s standpoint on the development of human knowledge. In qualitative 

research, these assumptions consist of ontology, epistemology, and axiology 

(Creswell, 2007). An ontological assumption questions the characteristics and 

nature of reality, and an epistemological assumption is based on the 

understanding of workings of the world (Creswell, 2007), whereas an axiological 

assumption questions the role of values in the research by combining both 

ontological and epistemological stances (Teherani et al., 2015). In this study, the 

realities (ontology) of RPA within the insurance industry are known, which include 

the benefits of RPA adoption to the insurers. However, the understanding of 

current drivers of RPA adoption and how technological, environmental, and 

organisational factors influence the adoption are unknown; hence, 

epistemological philosophy was deemed appropriate for this study. Epistemology 

allows the researcher to understand what is known about the adoption of RPA in 

the South African insurance industry and how the known factors influencing the 

adoption of RPA are perceived as positive or negative by the participants of this 

study.  

According to Sreeja and Shanthini (2018), the emergence of RPA has disrupted 

the insurance market. This automation technology has revolutionised core 

insurance business operations; however, the adoption of RPA presents both 

opportunities and challenges for insurers (Akkor & Ozyuksel, 2020; Lamberton et 



36 

al., 2016; Lanfranchi & Grassi, 2022). Given that there are varying views on the 

impact of RPA in the insurance market, an interpretive approach is used in this 

study, which is one of the epistemological stances. Interpretivism lends itself to 

discovering insights about a new phenomenon from an individual’s perspective. 

It is also found to be flexible enough to explore different explanations of the 

research problem, thus allowing the study to refine the focus into a narrower 

context as the research progresses (Saunders & Lewis, 2018). The reasons for 

selecting interpretivist philosophy are that it is compatible with qualitative 

methods of collecting data (Saunders & Lewis, 2018). Furthermore, the chosen 

philosophy enabled this study to seek multiple views on understanding the 

complexities of this topic, which included uncovering the underlying drivers 

(antecedents) of technology adoption at the organisational level as well as the 

contextual factors influencing the RPA adoption process within the South African 

insurance industry.  

This study is framed with assumptions influenced by the characteristics of 

qualitative research (Kothari, 2004). While inductive forms of reasoning are often 

associated with qualitative research, which involves drawing from a set of multiple 

premises (observations and data findings) and combining them to formulate a 

general theory (Saunders & Lewis, 2018), this study used abductive reasoning, 

which refers to the iteration between empirical data and existing theories found 

in the literature (Saunders & Lewis, 2018). Since the purpose of this study was to 

utilise existing theories as a guideline for the data collection process as opposed 

to generating a new theory from the data collected, the abductive reasoning 

approach is deemed appropriate. This approach allowed for this study to identify 

contextual factors influencing technology adoption from existing research and 

technology adoption theories whilst using empirical data to refine contextual 

factors emerging from analysis to better understand the RPA adoption process 

from an insurance perspective.  

While many studies addressing the adoption of RPA at an organisational level 

have used a case study design to test the factors influencing the adoption of 

technology based on a single adoption model (Flechsig et al., 2022; Holmberg & 

Härning-Nilsson, 2020; Pillay & Njenga, 2021), an exploratory design was 



37 

selected for this study. This research design focuses on understanding and 

gaining new insights on the topic through the integration of adoption models and 

empirical data from those who have interacted with RPA within the insurance 

landscape. The exploratory research design provides an opportunity to generate 

new ideas on a particular topic (Kothari, 2004), hence it was deemed to be 

appropriate for this study, as the study attempted to uncover the antecedents 

behind RPA adoption in the insurance industry and address the different 

perceptions of factors influencing RPA adoption based on the individual’s 

experiences.   

Based on the timeframe and limited financial resources allocated for the 

completion of this study, a cross-sectional timeframe was chosen for the data 

collection. Cross-sectional research is defined as a study of a specific topic 

conducted over a single period, i.e., a “snapshot”, and allows for the analysis of 

multiple variables (Saunders & Lewis, 2018). Additionally, the use of cross-

sectional methods allowed this study to conduct a quick and inexpensive data 

collection process. The study used semi-structured interviews and secondary 

data in the form of industry documents to identify the antecedents of RPA 

adoption and ascertain the perceived influences of the various factors.  

3.2 Unit of analysis  

This study understood that the adoption process of the RPA within the insurance 

industry involves both insurance organisations and the RPA vendors; therefore, 

the two units were considered for the analysis. The study identified senior 

managers within insurance organisations and RPA vendor companies. The 

reason for the two units of analysis was to allow the study to obtain perspectives 

from those with a direct strategic influence on the decision to adopt RPA within 

an organisation. It was also important for this study to provide insights from both 

sides (suppliers and adopters) involved in the RPA adoption process. This 

allowed for the richness in data and participant enrichment of the study.  

 



38 

3.3 Population 

Saunders and Lewis (2018) define the research population as all the elements, 

people, objects, events and services of a particular group, organisation, or 

universe. The target population refers to a population with specific interests in the 

research topic. This may include individuals or organisations that meet the 

sample criteria of the researcher (Klopper, 2008). For this study, the target 

population included insurance organisations that have adopted RPA, senior 

technical and business managers within the insurance industry, and RPA 

vendors that have procured RPA services from insurers. However, this study 

found it impossible to identify all the senior managers that are directly involved in 

the RPA adoption within the insurance industry, hence the sampling frame for this 

study could not be determined. A sampling frame is referred to as a source list 

consisting of all members of a targeted population (Saunders & Lewis, 2018). 

3.4 Sampling methods and sizes  

Sample designs can be categorised into two types, viz., probability and non-

probability sampling. Saunders and Lewis (2018) describe probability sampling 

as a technique used for the random selection of a sample from a known sampling 

frame, whereas non-probability sampling relates to a technique used when the 

sampling frame is unknown or inaccessible to the researcher. For this study, non-

probabilistic sampling is due to the inaccessibility of the sampling frame. Initially, 

participants were carefully selected based on purposive sampling. Purposive 

sampling is also referred to as subjective sampling, and it is based on the 

researcher’s judgement when selecting the participants of the study (Saunders & 

Lewis, 2018). However, due to low response in the predetermined sample list 

during the pilot stage of the study, the snowballing sampling method was also 

employed for participant enrichment. Snowballing sampling also forms part of the 

non-probability sample technique, in which the initially identified participants of 

the study will provide referrals and recommendations of potential candidates that 

can form part of the study (Saunders & Lewis, 2018). Therefore, this study made 

use of both purposive and snowballing sampling methods in the selection of the 

participants. The selected sample was found to be crucial in addressing the 



39 

research questions as all the participants were chosen based on the following 

characteristics. 

a) Participants should have similar RPA experience – all participants were 

selected based on their involvement in the RPA adoption process. 

b) Homogeneity – All participants occupied senior management positions 

(similar level of occupation).  

This study also employed some degree of heterogeneity in the sample 

characteristics. Participants should belong to an organisation, which could be 

either an insurance organisation or an RPA company only. RPA vendors were 

chosen to provide information from a supplier or technology owner context 

whereas insurers were to provide information from an adopter or user context 

based on their interaction with RPA. The approach of applying some degree of 

heterogeneity was to provide minimum variation in the data collected, thus 

generating deeper insights into the factors influencing RPA adoption in the 

insurance industry.  

According to Creswell (2007), a sample size of between five (5) and twenty-five 

(25) participants is recommended for semi-structured interview data collection 

methods. Due to time constraints, in this study a sample size of a total of twenty 

(20) sources of information was selected, which comprised seventeen (17) 

participants and three (3) industry documents. However, only twelve (12) 

participants were interviewed out of the seventeen (17) participants targeted for 

this study, which results in a 70.59% response rate. The response rate for this 

study is found to be suitable and aligns with the recommendations for semi-

structured interviews by Creswell (2012). The demographics of the participants 

involved in the study are provided in Table 3-1, and any identifiable information 

obtained during the interviews remained anonymous.  

In addition to the sample for the semi-structured interviews, three (3) industry 

documents were selected for analysis. The selection of these documents was 

limited to published reports from the three commonly used RPA vendors such as 

UiPath, Automation Anywhere and Blue Prism, which have partnered with South 

African insurers on their RPA adoption journey.  



40 

Table 3-1: Participants of the study  

Participant ID 
Industry 

Classification  
Role 

RPA Adoption Process: 

Involvement 

Participant A Insurance Automation Head  Decision Stage  

Participant B RPA Vendor 
Divisional Head: RPA 

Implementation  

Knowledge & Decision 

Stage  

Participant C RPA Vendor 
Senior Business Process 

Engineer  
Piloting Stage 

Participant D RPA Vendor 
Head Architect: Intelligent 

Automation  

Knowledge & Piloting 

Stage 

Participant E Insurance  
Business Intelligence 

Manager: Underwriting  
Piloting & Decision Stage  

Participant F  Insurance  
Executive Head: Insure 

Distribution  
Decision Stage  

Participant G RPA Vendor 
Business Development 

Manager 

Knowledge & Decision 

Stage  

Participant H RPA Vendor 
Implementation Partner: RPA 

Developer 
Implementation Stage  

Participant J  Insurance Senior BI Developer 
Piloting & Implementation 

Stage 

Participant K  Insurance  
Lead Architect: Process Re-

engineering  

Decision & Implementation 

Stage  

Participant L  Insurance  Digital Program Manager  
Knowledge & Decision 

Stage  

Participant M  Insurance  Data & Analytics Manager  Implementation Stage 

 

This study also acknowledges that for purposive sampling the size of the sample 

is dependent on theoretical saturation. Theoretical saturation, also known as data 

saturation, refers to a point during the data collection process where little or no 

new information can be found to generate new themes to address the research 

questions (Glaser & Strauss, 1967, as cited in Guest et al., 2020).  

 



41 

3.5 Data collection process 

The data collection procedure followed a series of steps aimed at collecting rich 

data to gain a deeper understanding of the factors influencing RPA adoption in 

the South African insurance industry.  

a) Development of the interview guide for the semi-structured 

interviews  

The interview questions are based on the factors identified in the conceptual 

framework and open-ended questions regarding the antecedents driving RPA 

adoption within the industry.  

b) Recruiting participants for the study.  

The recruitment strategy began with recruiting participants through professional 

connections, such as the Insurance Institute of South Africa (IISA). However, with 

regard to recruiting RPA vendor participants, engagements with the gatekeepers 

of those companies were made to obtain consent to conduct interviews. 

Gatekeepers refer to people in administrative positions such as HR or secretaries 

(Kabir, 2016). In addition, a social platform like LinkedIn was used to reach 

potential participants, followed by a formal request via email.  

c) Interviewing the participants  

Upon obtaining informed consent from participants, interview guidelines were 

sent to the participant before the interview session. This allowed participants to 

acquaint themselves with the topics to be covered. All interviews were conducted 

in a digital setting via Microsoft Teams, with a time allocation of 30-45 minutes 

per interview. The sessions allowed for follow-up questions to be asked by the 

participants and an open discussion to cover any additional factors the 

participants found had influenced the adoption of RPA in their organisation.  

d) Documenting the interviews  

Interviews were recorded via the Microsoft Teams platform and transcribed 

thereafter.  



42 

3.6 Research instrument  

The research instruments are supported by the research approach and design 

for this study, and semi-structured interviews and documents were chosen for 

this study.  

3.6.1 Semi-structured interviews 

The semi-structured interviews were guided by open-ended questions derived 

from the analytical framework and literature of other similar studies that explored 

the adoption of RPA in the financial services sector. Semi-structured interviews 

were selected because they ensure that rich information is captured from the 

interviewee since the conversation is not fully structured (Kothari, 2004). This 

merit ensured the gathering of rich and comprehensive data from the research 

participants. Furthermore, semi-structured interviews enabled the participants to 

engage in lines of conversation following themes but simultaneously allowing a 

degree of flexibility in responses from the participants. 

Following Kabir's (2016) semi-structured interviewing method, this study made 

use of an interview guide that consisted of a list of questions to be covered during 

the interview sessions. The interview guide questions were based on the factors 

identified under the technological, organisational and environmental contexts of 

the conceptual framework (see Figure 2-5). Furthermore, the questions of the 

interview guide aligned with the research questions of this study as shown in 

Table 3-2. 

Table 3-2: Consistency table (interview questions and research questions)  

Research Questions (see Chapter 

1.4)  

Interview questions (see APPENDIX 

C: Research Instrument)  

RQ.1. What are the potential 

antecedents to RPA adoption in the 

South African insurance industry? 

Question 2. Antecedents of robotic 

process automation adoption (include 

questions 2.1. and 2.2.) 



43