The Role of RPA and Data Analysis in the Transformation of the Insurance and Banking Industries

Abstract

Robotic Process Automation (RPA) is a software-based technology that uses configurable algorithmic software agents (bots) to replicate manual user activities across digital systems. It represents an evolution from earlier workflow scripting tools, and is distinguished by its ability to be used without requiring substantial IT infrastructure modifications or extensive programming knowledge. In the banking and insurance sectors, organizations face increasing pressure to adopt modern technologies that streamline operations and reduce costs while complying with strict regulatory requirements. Robotic Process Automation (RPA) has emerged as a viable and cost-effective solution, enabling automation of repetitive and rule-based tasks without requiring major changes to legacy IT systems. This paper conducts a literature review to examine the current use cases of RPA technologies in banking and insurance, analyzing how these technologies are employed to enhance corporate efficiency and performance. The review draws from recent academic publications and case studies between 2017 and 2025, identifying core implementation areas such as customer onboarding, claims processing, compliance reporting, and underwriting automation. The results highlight substantial improvements in processing speed, error reduction, and resource optimization, along with evolving metrics for measuring effectiveness. The study concludes by identifying key success factors, performance measurement approaches, and challenges in RPA implementation, offering insights for both practitioners and researchers aiming to understand the role of automation in financial services transformation.

1. Introduction

Today’s organizations are experiencing unprecedented change driven by emerging technologies and digital transformation initiatives. Amidst the emergence of AI, ML, and advanced analytics, Robotic Process Automation (RPA) has become a pivotal technology evolving from traditional scripting languages, and enabling organizations to re-establish their operational models with a lower technological burden and lower needs for highly specialized IT and Software Engineering teams. RPA represents a form of business process automation (BPA) that utilizes software robots or “bots” to emulate human interaction with web applications and desktop-based software, effectively automating repetitive, rule-based tasks without requiring fundamental changes to existing IT infrastructure [1,2]. The financial services industry, particularly the banking and insurance sectors, has been at the forefront of RPA adoption, not only because of the general drive for optimizing operational efficiency that is shared across all industries, but more importantly due to the distinctive characteristics of their work: large volumes of repetitive, rule-based transactions, substantial volumes of structured data, a high dependency on legacy IT systems, and stringent compliance obligations. These factors make RPA especially well-suited for automation in these contexts [3].

This literature review aims to address two critical questions: What are some of the specific applications of RPA technology in the insurance and banking sectors, and how is corporate performance measured? This includes important questions such as how this technology is being evaluated in regard to its effectiveness, the results achieved, and the benefits gained, as well as some new challenges faced, such as deciding when it makes more sense to choose different options or even potentially consider decommissioning automated robots for alternative solutions. This review intends to provide insight into how companies in these industries leverage RPA as a technology and not to critique the previous work of other studies directly. Considering the rapid evolution of RPA technology, prior work and current implementations may differ, and direct comparison between the two may therefore not be possible on equal terms.

Successful RPA implementation depends on several key considerations: comprehensive process assessment for identifying optimal automation candidates, change management initiatives addressing workforce concerns, and governance frameworks measuring effectiveness against strategic objectives [4]. These factors are shown in Figure 1.

This research aims to present recent scholarly findings on RPA applications in banking and insurance, analyze methodologies for measuring performance improvements, and identify best practices for maximizing automation benefits. It is anticipated that this exercise will contribute valuable insights for stakeholders, such as financial services executives, technical leadership, and groups that take part in digital transformation initiatives, such as professional services consultants, client advisors providing services related to process automation, and academic researchers investigating business process technologies, information systems, and business analytics.

2. Materials and Methods

The selected literature reveals several prominent themes regarding RPA applications and performance measurement in the banking and insurance sectors. These themes are analyzed below with reference to significant contributions from various researchers.

2.1. RPA Applications in Banking

In their comprehensive study of RPA implementation in global banking operations, Lacity and Willcocks [3] identified multiple high-value use cases, including account opening and maintenance, loan processing, payment reconciliation, and compliance reporting [3]. Their research, based on 16 case studies across European and North American banks, demonstrated that RPA deployment reduced processing times by an average of 43% while improving accuracy rates by 38% [3]. The authors emphasized that successful implementations were characterized by cross-functional governance structures and clear performance metrics established during project initialization [3].

Moffitt et al. [4] conducted an in-depth analysis of RPA deployment in customer onboarding processes [4]. The study revealed that automated onboarding reduced processing time by up to 87.5% while decreasing compliance errors by 32% [4]. The research further observed that banks leveraging RPA for automation for auditing procedures (including know-your-customer (KYC) procedures) can also reduce compliance-related errors, demonstrating the dual benefits of operational efficiency and risk mitigation [4].

Research by Aguirre and Rodriguez [5] focused on RPA applications in middle- and back-office banking functions. Their study of eight Spanish banks found that automation of data reconciliation processes yielded average cost savings of 35% while reducing full-time equivalent (FTE) requirements by 28% [5]. The authors identified critical success factors, including process standardization prior to automation and integration of performance measurement systems with existing key performance indicators (KPIs).

Thekkethil et al. [6] present RPA use cases of loan processing, customer information collection, loan monitoring, loan approval, and automatic loan pricing. The study also highlights the use of RPA for fraud and threat detection. RPA technology is employed to identify and counter fraud based on multi-source data extraction. Using rule-based software bots, leading banks and financial institutions (based in India) have significantly benefited from switching away manual tasks from humans and successfully employing software bots which can run 24/7 reducing errors, increasing the overall value of the involved banking services [6].

Rane et al. [7] researched the evolving capabilities provided by the use of artificial intelligence (AI), natural language processing (NLP), and RPA in the domain of corporate finance. The research paper focuses on the transformative capabilities of RPA in finance corporations as an enabler for digital transformation, innovation optimizing resource allocations and promoting sustainable development. The study highlights the rapidly increasing automation possibilities of routine tasks using RPA and making data-informed decisions combining the innovative methods by introducing NLP, AI, and RPA into traditional finance methodologies [7].

The systematic review of Patrício, Varela and Silveira [8] further focuses on the ever-increasing significance of combining artificial intelligence (AI) and Robotic Process Automation (RPA), which can pose both challenges and opportunities. The study proposes a model that also takes into consideration social, environmental, and economic factors and attempts to achieve a balance between the business-related benefits and the multifaceted implications that are part of every technology evolving responsibly. The study highlights the absence of methodologies and the fact that AI and RPA remain relatively underexplored, specifically in relation to their economic outcomes versus their social and environmental impacts [8].

Ayinla et al. [9] reveals significant benefits in terms of operational efficiency, data processing accuracy, and regulatory compliance. The study emphasizes the need for modern-day accountants to prepare for a new iteration of advancements in the digitalized environment. Based on the recent developments in AI, predictive analytics, decision support systems, and RPA, the study anticipates new challenges that will need to be managed effectively. Continuous learning combined with effective strategy- and policy-making will be pivotal for the ethical use of modern-era technology in accounting practice [9]. Finally, the study stresses the need for further studying of the long-term impacts of AI and RPA technology in revitalizing the role of accountants, which are not fully clear as of yet. The RPA applications in Banking are summarized in

Table 1. RPA applications in banking.

Process Categories	Examples
Account opening	Ticket-based procedures, system data-entry, etc.
Loan processing	Loan approval, loan monitoring, loan pricing
Payment reconciliation	-
Compliance reporting	Reporting procedures, due diligence tasks, information collection from various sources, form-based automation, etc.
Customer onboarding (customer information collection)	Data collection, data processing and entry
Know-your-customer (KYC)	Reporting, searching various data sources, etc.
Middle- and back-office functions	Data extraction, document extraction, OCR, data entry, other ticket-based processes
Accounting/Finance	Data reconciliation, corporate finance, data entry, data collection
AI and ML	Decision-making, sentiment analysis, NLP
HR task automation	Employee onboarding/offboarding processes, sending notifications, email automations (mandatory training notifications, etc.)
Other office-related tasks (sending reminders/notifications, office booking automations, etc.)	Reporting, processing forms, reserving office spaces, data entry, data extraction, etc.

below.

2.2. RPA Applications in Insurance

The insurance sector presents distinctive RPA applications, as documented by several researchers. Lamberton et al. [10] examined automation initiatives across 12 insurance companies, identifying claims processing as the primary RPA use case. Their findings indicated that automated claims handling reduced processing time by an average of 50% while decreasing processing costs by 30–40% [10]. The study further noted that insurers implementing RPA in claims functions reported 22% higher customer satisfaction ratings compared to pre-automation benchmarks [10].

Anagnoste S. [11] investigated RPA deployment in underwriting operations across property and casualty insurers. The research demonstrated that automation of data collection and validation tasks within underwriting workflows reduced policy issuance time by 60% while improving data accuracy by 67% [11]. It also highlighted that performance gains were most significant when RPA implementation was accompanied by process standardization and elimination of unnecessary process variations [11].

A significant contribution to understanding RPA applications in insurance compliance functions was made by Hofmann et al. [12]. Their study of regulatory reporting automation among seven European insurers revealed that RPA reduced compliance resource requirements by 35% while accelerating reporting cycles by 40% [12]. The authors emphasized that automated compliance processes demonstrated 98.5% accuracy compared to 92% for manual processes, representing a substantial risk reduction benefit beyond efficiency improvements [12].

Kumar, M. [13] in their study, focus on the automation of time-intensive administrative tasks, including claims processing, policy renewals, customer service processes, data entry, inventory management, and billing and payments. The study also includes insights regarding RPA-powered chat bots and virtual assistants and AI-enabled capabilities which expand the range of applications of RPA in the health insurance sector [13]. The study presents benefits including faster responses enabled via chatbots, a reduction in errors, time savings, minimization of delays through the employment of automated reminders, improvement in customer engagement, a reduction in claims processing time, and improved accuracy.

Mahadevkar et al. [14] study how traditional RPA application in insurance can be further improved by using AI and Generative AI (GenAI) [14]. The study emphasizes how AI can extend the range of RPA applications to work more effectively with unstructured or semi-structured data (an area where RPA technology typically struggles) and highlights the new capabilities that AI can offer.

Chaturvedi and Sharma [15], in their review, anticipate that the use of RPA in healthcare will increase in the future, given the continuous advancement of technology [15]. The study also presents future directions of RPA expansion in healthcare, including a. integration with AI and machine learning, b. addressing integration challenges, and c. further expanding RPA applications to new domains where, typically, no RPA solutions are readily available. However, the study also presents some ethical considerations, primarily regarding the automated processing of patient records and the need for regulatory measures regarding privacy, data security, and patient confidentiality [15]. The RPA applications in insurance are summarized in

Table 2. RPA applications in insurance.

Process Categories	Examples
Automated claims processing	Ticket-based procedures, system data entry, etc.
Underwriting operations	Property and casualty insurance sectors
Automated data collection and validation	-
Regulatory reporting automation/automated compliance processes	Generating reports, due diligence procedures, information collection from various sources, form-based automation, etc.
Customer onboarding (customer information collection)	Data collection, data processing and entry, document processing
Back-office functions	Automation of processes in terminal emulators, reporting, searching various data sources, etc.
Commercial operations	Data extraction, document extraction, pricing monitoring/comparison
Accounting/finance	Data reconciliations, corporate finance, data entry, data collection
Marketing operations	Sending promotional emails, management of campaigns, automated management of potential customers, automation of various forms
HR task automation	Employee onboarding/offboarding processes, sending notifications, email automations (mandatory training notifications, etc.)
Other office-related tasks (sending reminders/notifications, office booking automation, etc.)	Reporting, processing forms, reserving office spaces, data entry, data extraction, etc.

below.

2.3. Corporate Performance Measurement Approaches

The literature reveals diverse approaches to measuring corporate performance in relation to RPA implementation. Cooper et al. [16] proposed a multidimensional framework categorizing RPA benefits into four domains: operational efficiency, cost optimization, quality enhancement, and value [16]. Their research across 23 financial institutions demonstrated that comprehensive measurement approaches incorporating metrics from all four domains correlated with 34% higher reported satisfaction with RPA investments compared to organizations focusing solely on efficiency metrics [16].

Syed et al. [17] conducted a longitudinal study of RPA performance measurement practices across 19 banks and insurance companies [17]. They identified three predominant measurement approaches: (1) cost-centric models focused on FTE reduction and direct savings; (2) process-centric frameworks emphasizing cycle time reduction and quality improvements; and (3) integrated balanced scorecards incorporating customer experience and employee satisfaction metrics [17]. Their analysis revealed that organizations employing integrated measurement approaches achieved 28% higher returns on RPA investments compared to those utilizing single-dimension metrics [17].

Research by Fernandez and Aman [18] examined the evolution of RPA performance measurement over implementation lifecycles. Their study of 14 financial institutions demonstrated that measurement emphasis typically shifted from cost reduction during initial implementation (88% of cases) to customer experience and strategic value during maturity phases (64% of cases) [18]. The authors argued that predefined measurement evolution paths were associated with more sustained RPA benefits compared to static measurement approaches [18]. Figure 2 summarizes the three performance measurement approaches.

2.4. Critical Success Factors and Implementation Challenges

Several studies highlight factors influencing RPA implementation success and performance outcomes (Figure 3). Santos et al. [19] identified five critical success factors based on 27 RPA implementations: (1) process standardization prior to automation; (2) clear governance frameworks; (3) stakeholder engagement and change management; (4) integration with existing systems; and (5) continuous performance monitoring [18]. Their research indicated that organizations addressing all five factors achieved 42% higher ROI compared to those focusing on technical implementation aspects alone [19].

Challenges in RPA implementation and their impact on performance outcomes were examined by Leopold et al. [20]. Their study of failed or suboptimal RPA initiatives identified common pitfalls, including inadequate process analysis (present in 76% of underperforming cases), a lack of IT-business alignment (68%), insufficient attention to exception handling (57%), and inadequate performance measurement frameworks (61%) [20]. The authors emphasized the necessity of comprehensive readiness assessments prior to RPA deployment. Horvat et al. [21], in their study, also recognize the importance of maintenance of the automations (for example, when underlying applications or user interfaces change), and the need for “change management” from both an organizational and a technical perspective [21]. Therefore, the human factor is of significant importance, not only in the development phase, but perhaps even more importantly in the post-production phase. Without proper maintenance and organizational support (such as from business users and stakeholders involved in the automation), the RPA project will have a short lifecycle and any benefits achieved will not be long-lasting. These challenges are summarized below in Figure 4.

2.5. Research Gaps and Opportunities

Despite the growing body of literature on RPA in banking and insurance, several research gaps remain. These gaps are summarized below (Figure 5). First, while numerous studies document immediate operational benefits, research on sustained performance improvements remains limited. Secondly, methodological approaches for isolating RPA impacts from concurrent process improvements or system changes are underdeveloped. Research on the integration of RPA with complementary technologies such as artificial intelligence, agentic automation, and cognitive automation represents an emerging area that has not been deeply investigated and requires further research through data analysis Pingili [22]. Also, some argue that RPA might be not be being used to its full potential. According to Mamede et al. [23], using it to its full potential would entail using process improvement techniques (such as a lean approach) before applying the automation itself. There are also further research opportunities. For instance, the term “microbots” has emerged in recent years as an evolution of Robotic Process Automation oriented towards automating specific parts of processes instead of automating entire processes (Naga [24]). In theory, the “microbots” approach focuses on modularity and reusability and represents a more light-weight approach. However, as of yet, the scope of actual implementations and use cases of this new architecture remain relatively obscure in the public space. Additionally, a general limitation of traditional RPA systems remains in their limited ability to process unstructured data, such as scanned documents, emails, and handwritten notes, due to their complexities Mahadevkar et al. [14]. Furthermore, while RPA can handle system-to-system transitions using structured data in such tasks, it requires support from artificial intelligence for more complex functions. Tasks such as scanning documents for information, comparing information with slight variations, anomaly detection, classification, and sentiment analysis are typically beyond the scope of traditional RPA-only solutions.

This literature review demonstrates the rising attention to various RPA applications and performance measurement in financial services contexts. The selected studies provide evidence of significant operational, financial, and customer experience benefits while highlighting the importance of structured implementation approaches and comprehensive measurement frameworks. This research aims to build upon this foundation by synthesizing these insights into an integrated understanding of RPA performance dynamics in banking and insurance environments.

2.6. Methods

This literature review employed a systematic approach to identify relevant research regarding the use of RPA technology in the banking and insurance sectors and approaches to measuring corporate performance in relation to RPA implementation. The data collection process was conducted using a search with keywords specific to RPA technology (

Table 3. Review search criteria used (shown as organized by category).

Categories	Keywords used
RPA	‘RPA’, ‘Robotic Process Automation’, ‘RPA technology’, ‘Software Automation’, ‘Low Code/No code’, ‘Low Code’, ‘No Code’, ‘RPA tools’
Insurance sector	‘Insurance’, ‘Insurance sector’, ‘Healthcare’, ‘Underwriting’
Banking sector	‘Banking’, ‘Banking sector’, ’Finance’, ‘Financial Services’, ‘Accounting’, ‘Corporate finance’
Performance metrics	‘Corporate Performance’, ‘KPIs’, ‘performance’, ‘performance metrics’, ‘evaluation’, ‘efficiency’, ‘savings’, ‘costs’, ‘benefits’, ‘issues’, ‘estimation’, ‘measurement’, ‘impact’

) (‘RPA’, ‘Robotic Process Automation’, ‘Low Code’, and ‘Low-Code/No code’, as well as related terms such as ‘Low-Code/No Code’, ‘Software Automation’, and others), financial services (using keywords such as ‘banking’, ‘finance’, and ‘financial services’), the insurance sector, and performance (‘performance measurement’, ‘evaluation’, and more). The search was performed using mainly Google Scholar and the following academic digital libraries: ScienceDirect, Springer, and scientific journals such as the Journal of Information Technology Teaching Cases, as well as other sources. The aforementioned sources are indicative and it should not be implied that these were selected specifically against other sources for any reason.

The publication time period was restricted to reports published from 2017 to 2025, corresponding to a significant period of rapid growth in RPA technology and its implementation in corporate financial and insurance environments. The language was restricted to English-only sources to ensure that no misinterpretation issues would occur due to the need to use automated translation tools.

The initial selection process was performed by examining the titles and abstracts of the identified publications, and those that were selected matched the following criteria:

1. Having a focus on RPA technology, rather than general references to digital transformation technologies and other tools such as entirely AI-driven solutions and cloud-based low-code/no code software, which, at present, are considered as intended to complement, and not entirely replace, RPA-driven automation.

2. Containing or focusing on the banking and insurance sectors or related to domains such as those of finance services and healthcare. For example, companies with business activity in e-Commerce, corporate banking, electronic payment providers, etc.)

3. Reporting use cases, evaluation methods and performance outcomes, application of performance measurement frameworks, and impacts and challenges or benefits of RPA technology implementation.

4. Documenting or containing discussion of results, case studies, analysis, and data-driven results.

In terms of geographic distribution, the range of studies reviewed was not limited to any specific region. The review included studies from North America, Europe, and Asia. No geographical restrictions were applied.

The methodology collected evidence seeking answers to the main and secondary questions that this research aims to explain, and concluded with the collection of material and documentation of the bibliography.

There are some limitations that need to be acknowledged. First, the reliance on the published literature may introduce bias due to possible over-representation of successful RPA implementations compared to problematic and/or failed implementations (which rarely are documented), and references to the latter are considerable rarer and difficult to find. Secondly, RPA technology is a relatively new technology that has rapidly involved in the past few years. Because of this, findings from earlier publications may no longer fully reflect current capabilities or practices (such as Process Mining, Agentic Automation, and recently added AI features, and more) that have changed dramatically over the past few years. Third, proprietary constraints and non-disclosure agreements (NDAs) for professionals involved in RPA (Business Analysts, RPA developers, major consultant firms, and other companies offering professional services) working on automation projects in corporations in this sector pose obstacles in gathering sources and detailed material (for example, interviews or questionnaires) and exact use cases. This may result in restrictions on the depth of available information, under-documenting of actual use cases, or substitution of precise use case details with more general information due to these conditions. While this approach reduces bias by relying on the above predefined criteria, we acknowledge that some case studies (including published, unpublished, or proprietary studies) may not be captured.

Despite these limitations, the methodology followed reveals important findings regarding the applications of RPA technology in the banking and insurance sectors, as well as different approaches to measuring performance outcomes. These have been organized and are presented by key thematic areas in the Materials and Literature Review Section. The findings derived are presented in the subsequent Results and Discussion Section below.

3. Results and Discussion

The systematic analysis of the selected literature reveals significant findings regarding RPA applications in the banking and insurance sectors and approaches to measuring performance outcomes. This section organizes these findings by key areas and discusses their implications for understanding the relationship between RPA implementation and corporate performance.

3.1. Prevalent RPA Applications in Banking and Insurance

The research reveals distinct patterns of RPA adoption across banking and insurance functions, with certain processes emerging as primary automation candidates based on their characteristics and potential returns.

3.2. Banking Sector Applications

In the banking sector, customer onboarding, account management processes (such as account opening and account closing), and compliance reporting (due diligence procedures, mandatory compliance requests from legal, tax and law enforcement authorities, etc.) represent the most prevalent RPA applications. These cases appear in the majority of baking-focused studies. These processes typically involve data extraction, collection, validation, and entry across multiple systems—tasks well-suited to RPA capabilities. As noted by Moffitt et al. [4], automated onboarding reduces processing time while decreasing compliance errors. Additionally, loan processing emerges as another significant application area. The research of Lacity and Willocks [3] showed that RPA implementation in loan processing reduced cycle times by an average of 43% and simultaneously improved compliance accuracy by 38%. This is attributed to the reduction or even elimination of manual data processing and to a reduction in errors due to the increased consistency of the RPAs which perform the same business steps with advanced data validation rules.

Regulatory compliance and reporting functions appeared in approximately half of the banking-focused studies. The implementation of RPA in organizational compliance has demonstrated particular value in various banking and finance firms. The main benefits reported were an overall improvement in the working experience for compliance teams and employees, a reduction in processing errors, particularly during the generation of reports, and the capabilities of RPA acting as a connecting layer between multiple systems (both between legacy banking systems and other more modern internal systems, and between internal banking systems and other external compliance tools). For example, RPA use cases reported include the automation of customer-specific data retrieval between the Lexis-Nexis suite of compliance and AML tools and various legacy banking systems that do not offer any APIs—meaning that RPA is most likely the only way to automate without requiring complex and often costly redesign of the entire software systems involved. Finally, another often overlooked advantage is the improved audit trail and logging capabilities that enterprise-grade RPA systems offer.

3.3. Insurance Sector Applications

In the insurance sector, claims processing emerges as the dominant RPA application area in the vast majority of insurance-focused studies. The research of Lamberton et.al. [10] demonstrated a reduction in processing time by approx. 50% compared to manual claims handling, accompanied by a decrease in processing costs by at least 30% or more [10]. However, these improvements were mostly for standardized, high-volume claims categories (for example: claims related to the motor sector), while other categories showed less gains [10].

The second major category is underwriting process automation. Hofmann et al. [12] found that the use of RPA typically applied to the data collection and validation components of underwriting workflows, explaining that these steps are fairly standardized and typically do not involve or require human judgment [12]. The degree of dependence of a business process on decision-making is a major factor that determines the feasibility of any RPA-based automation (at least those that do not depend on AI features, etc.).

Another application of RPA are automations for customer care and customer service departments in insurance firms. Though several reports mention these cases in general (such as data migration, system integration, automated reporting, etc.), many of these studies do not document any specific use cases. This review found that these cases were not very well documented. As these cases often are business-specific use cases, their scope is generally difficult to present in a study due to their references to in-house business terminology and business logic components and business rules which typically are not shared across different organizations and can differ vastly between them. Additionally, they are often categorized as cross-sector applications, meaning that they do not belong specifically to either the insurance or banking sector; for instance, RPA use cases in Human Resources (HR). Finally, the limited source of information may be a contributing factor to the lack of specific use cases, as well as the fact that they often are not publicly available due to corporate data restrictions and NDAs which prohibit the publication of use cases, as has been stated in the Limitations paragraph of the Methods Section. This is a commonly known fact in RPA and the software industry in general. However, it is important that these cases are mentioned, and for this reason, they have been included in this section.

3.4. Regarding Performance Measurement

The analysis reveals diverse approaches to measuring RPA performance outcomes, with organizations employing both quantitative and qualitative metrics. Admittedly, the predominant measurement is financial performance-related metrics.

Cost reduction metrics: These are the most frequently reported performance indicators appearing in the entire literature. These metrics include labor cost savings (the number of manual work hours saved, FTE reduction or reallocation). It is worth mentioning that frequently, the number of hours saved might not be represented accurately, and may be either exaggerated or underestimated. This happens due to the fact that it is difficult to estimate the exact amount of time required, which may vary per case, and also due to the fact that sometimes the automation may perform extra checks (such as data validations, or waiting for the user interface (UI) of an application to load or appear), all of which the users may not always be in a position to estimate.

Return on investment (ROI) calculations: These appear in almost all of the studies. The typical time period for ROI calculation ranged between 6 and 24 months). The studies revealed that organizations may apply variations in regard to measuring ROI. The research revealed key factors being business case methodologies, the level of maturity of the RPA implementations, the complexity of the automations’ scope, etc. For example, some companies also estimate cost avoidance as a factor. This is common in automations developed for compliance functions, where these operations are exposed to regulatory penalties, especially if reporting is not accurate or delivered on time to the involved authorities. For some organizations, this is also quantified and represented in the end ROI calculations Leopold et al. [20].

Operational Performance Metrics: Processing time metrics are also employed in the studies. The literature consistently reports significant time savings based on calculating the average processing times and performing a comparison between the time spent by people versus the time that robots spend to process the same amount of data. Error reduction metrics are also referenced by most studies. For example, the average error rate reduction is a key metric. Error rates are frequently split into different error types depending on the implementation. Most commonly, errors are split business-related errors (for example, missing data or errors during data validation, business-specific rules, or conditions not being met), application-related errors (such as the underlying application not loading or displaying data properly, etc.), and RPA-specific errors (such as not being able to locate a specific window or field, failing to click on a particular element, or failing to extract data from). These are all errors that most organizations monitor and try to understand the origin of, and try to potentially optimize their implementations based on this.

Customer Experience Metrics: Based on the overall study of the literature, customer satisfaction measurements are frequently employed, with nearly half of the studies mentioning some form of a measurement of satisfaction. For example, insurance customers whose claims were processed through automated workflows reported an increase in the Net Promoter Score (NPS) based on satisfaction survey methodologies according to Lamberton et al. [10]. Additionally, response time metrics appear in most customer-facing processes. For example, RPA implementation in banking onboarding reduced customer waiting times by reducing the time taken from weeks to days or even hours, representing a significant improvement in service reliability.

Employee Impact Metrics: Job satisfaction and engagement metrics appear in some of the studies. The main reason for improved employee satisfaction cited is the improved quality of daily workload, a reduction in manual repetitive tasks, and a reduction in mundane tasks, allowing workers to focus on more creative tasks involving critical thinking (Cooper et. al, [16]). For organizations, the transition of workforce activities translates into benefits mostly in terms of employee retention, improved employee engagement, and the achievement of workforce transition strategies. RPA-driven digital transformation, according to Vilaplana et al. [25], represents a change in both employer and employee mentalities. Shet and Pereira [26] emphasize the role of managers as paramount in motivating individuals and maintaining collaboration within teams. Gandía et al. [27] agree with the notion that the influence of digital skills in society and their implications in the business area need to be further investigated.

Other Measurements: Both the banking and insurance sectors are constantly seeking cost-effective services to gain customers’ trust and exploring ways to create long-lasting relationships with their customers as some of their core objectives. Though these outcomes cannot be directly measured, Rizvi and Srivastava [28] point out that RPA not only enables these organizations to achieve these objectives, but also helps them to withstand growing competition and evolution [28]. Therefore, other metrics, including indices relevant to the organization’s level of technological competitiveness and innovation, as well as benchmarking analysis to compare the organization against similar companies, can be explored further.

3.5. Discussion

While prior studies consistently report efficiency gains from RPA adoption, there are notable differences in scope, methodology, and measurement approaches. For example, Lacity and Willcocks [3] reported a 43% reduction in processing times across banking operations [3], whereas Aguirre and Rodriguez [5] found cost savings averaging 35% in Spanish banks, though their analysis emphasized FTE reduction rather than cycle time. These variations highlight differences in focus between time savings versus labor savings in their reported outcomes.

In terms of compliance automation, Hofmann et al. [12] demonstrated 98% accuracy in automated regulatory reporting tasks, compared to 92% for manual processes [12], whereas Moffitt et al. [4] focused on onboarding and auditing, also showing a reduction in compliance errors, but offering less discussion of long-term sustainability [4]. Together, these findings suggest that though, overall, RPA consistently improves compliance-related processes, the reported benefits can vary regarding whether the evaluation prioritizes error reduction, time efficiency, or risk mitigation.

In insurance claims processing, Lamberton et al. [10] reported a 50% reduction in processing time across standardized claims, while Anagnoste [11] emphasized underwriting automation, highlighting improved accuracy (67%) rather than time savings [11]. These studies illustrate that the type of process automated (standardized vs. knowledge-intensive) determines the primary benefit reported. This comparison shows the need for more complete evaluation frameworks that consider both efficiency and quality improvements.

Finally, a limitation shown across multiple studies is reliance on short-term performance metrics. Fernandez and Aman [18] noted that the measurement emphasis typically shifts from initial cost reduction at the early adoption phase to customer experience and strategic value in later phases [18]. This aligns with other studies like that of Syed et al. [17], who observed that organizations employing integrated performance frameworks achieved higher returns compared to those using cost-centric models alone [17]. This suggests a gap in the literature: few studies systematically track RPA performance across the implementation lifecycle, limiting a more thorough understanding of its long-term impacts. In general, the need for more focused research (particularly on performance evaluation methodologies, as well as on the overall impact of AI and RPA in the economy and society) seems to be something that some of the studies reviewed seem to converge on [8,17,22,27].

4. Conclusions and Prospects

This study of the literature and case studies provides a means to understand the extent of the positive effects of RPA on organizations, particularly in banking and insurance. The study concludes with confirmation of the significant impact that RPA has made on the efforts of organizations and the benefits that the technology has generally provided in terms of helping employees to perform certain tasks. While RPA technology provides the opportunity for organizations to harness its benefits to increase productivity and efficiency, it also faces several challenges in implementation and utilization (such as a lack of readiness, scaling difficulties, low ambition among top executives, or even ethical considerations related to workforce implications). It is worth noting that despite the adoption of RPA, data analysis and reporting of data resulting from RPA use remain relatively underutilized.

In regard to challenges, another point needs to be highlighted. The deployment of RPA solutions should not be understood as something that only raises concerns about job displacement and widespread unemployment. RPA is, and will be more so in the coming years, an artificial partner that enhances workforce capabilities, as well as business outcomes. In this sense, a social and ethical challenge arises regarding effective human–machine collaboration in data management. This collaboration, which manifests itself within the digital field, could change the values and personal traits of humans (Tzavaras & Stelios [29]). Let us not forget that intelligent automation technologies, although distinct from human intelligent actions, are often used interchangeably (Sgantzos et al. [30]), gradually closing, at least at a conceptual level, the gap between a natural and an artificial agent. Everything begins with a human command, and future challenges should not be exclusively regarded as only economic in nature (see f.i. unemployment). Automation represents still a human-based choice that can reshape the inherent foundations of our moral universe. The importance of the human factor across the entire lifecycle of an RPA project is also recognized in most studies as a key consideration, and for some, many aspects involving humans pose important challenges (i.e., development, maintenance, interaction with business users, adaptation to change, etc.) [19,20,21].

During our research, it became evident that most publications relating to the finance sector primarily address banking institutions, while other financial institutions are significantly less popular. This observation appears to be in accordance with other recent studies which examine data from recent years (Horvat et al. [21]).

It is also evident that upper management in banking institutions are starting to take initiatives to increase the level of involvement of non-technical business users, who, through training, are expected be able to create their own automations. This has led to an increase in the range of potential use cases that can be automated without the direct involvement of IT departments and/or professional developers. This coincides with a concept that many RPA companies (such as Microsoft UiPath and others) refer to as “the democratization of Robotic Process Automation”. This also relates to their efforts to incorporate new features into their software products and to prove that business process automation is no longer the sole domain of IT departments or specialized consulting companies and software developers. Instead, virtually every employee can achieve significant results through the use of low-code/no-code tools. In practice, this means (at least for the majority of the cases) making it possible for business users to leverage RPA to create their own attended automation workflows (starting with simple processes before advancing to more complex cases), provided that they have received structured training programs and are given ongoing support in effectively utilizing RPA platforms.