Economics and Business
Quarterly Reviews
ISSN 2775-9237 (Online)




Published: 18 November 2026
Fraud Detection Beyond Technology: The Role of Auditor Agility
Ahmad Mulia Mediansyah, Harnovinsah, Syahril Djaddang
Pancasila University

Download Full-Text Pdf
10.31014/aior.1992.09.04.735
Pages: 52-69
Keywords: Auditor Agility, Information System, Whistleblowing, Professional Skepticism, Fraud Detection
Abstract
This research was conducted to analyze the impact of information systems, whistleblower systems, and professional skepticism on fraud detection, as well as explore the role of auditor agility moderation. The approach used in this study is a quantitative approach, data that was successfully collected by 197 internal auditor respondents in Indonesia and analyzed through Partial Least Squares–Structural Equation Modeling (PLS-SEM). The results of this study revealed that information systems and whistleblower systems did not significantly affect fraud detection, suggesting that both mechanisms tended to function primarily as operational and compliance support tools rather than strategic fraud detection instruments. Conversely, professional skepticism and auditor agility significantly increase the effectiveness of fraud detection. In the role of moderation, the agility of auditors strengthens the relationship between information systems and fraud detection, but does not moderate the relationship involving the whistleblower system and professional skepticism. These findings highlight that the effectiveness of fraud detection is not only determined by the existence of formal systems or technological infrastructure, but is highly dependent on auditors' adaptive and cognitive abilities. This study contributes to the fraud detection and audit literature by expanding on Dynamic Capability Theory and demonstrating that auditor agility serves as a selective dynamic capability that enhances the effectiveness of technology-based mechanisms in audits.
1. Introduction
Fraud refers to illegal actions committed by individuals or groups for personal gain at the expense of other parties (Khofidah Pujilestari & Nera Marinda Machdar, 2024). Fraud has become a major challenge for both public and private organizations because it not only causes financial losses but also damages organizational reputation and reduces stakeholder trust, which in some cases may threaten organizational sustainability (Rahmawati et al., 2021).
As a result, fraud detection has become increasingly important for organizations and companies. The Report to the Nations 2026 published by the Association of Certified Fraud Examiners (ACFE) (ACFE, 2026) shows that asset misappropriation remains the most common type of fraud, accounting for 53% of reported cases, followed by corruption at 8%. Although financial statement fraud represents only 1% of cases, it results in the highest financial losses. The report also reveals that 38% of perpetrators commit more than one type of fraud simultaneously, indicating that fraud schemes are becoming more complex and require more adaptive detection approaches.
The same ACFE report also shows that Indonesia ranked third in the highest number of reported fraud cases by country/territory in the Asia-Pacific region, with 20 cases (13.99%) out of a total of 143 cases, following Australia and China. This finding suggests that existing fraud prevention and detection mechanisms have not yet been fully effective. In addition, the ACFE reported that only 15% of organizations were able to recover their losses in full, while 29% recovered only part of the losses, and 56% were unable to recover any losses at all. These findings underscore the importance of effective fraud detection in minimizing organizational losses and preventing further damage.
Several major cases in Indonesia further highlight this urgency. In late 2024, e-Fishery was reported to have manipulated its financial statements through a dual reporting system that allegedly inflated revenue by nearly USD 600 million, with around 75% of the reported figures suspected to be invalid. In another case, early 2025 investigations revealed alleged corruption related to crude oil import governance involving subsidiaries of Pertamina, with estimated state losses reaching IDR 193.7 trillion. These cases demonstrate that weak fraud detection systems and ineffective organizational responses to fraud indicators can lead to significantly greater financial and reputational damage.
Along with the growing complexity of business activities and the increasing digitalization of organizational processes, fraud detection can no longer rely solely on traditional audit approaches. Information systems have become an important tool in supporting fraud detection by providing integrated, accurate, and real-time data that help auditors identify unusual patterns and potential fraud indicators at an earlier stage (Adam, 2023).
Several previous studies have found that information systems significantly influence fraud detection (Damayanti & Nugroho, 2024; Najatu Silva & Aufa, 2023; Nurhasanah et al., 2019). However, other studies reported insignificant results (Choirunnisa, 2022; Meiryani, et al., 2023; Polontalo et al., 2022). Conclusion These non-uniform findings suggest that the effectiveness of information systems is determined not only by the quality of the technology itself, but also by the auditor's ability to utilize and respond to the information generated by the system.
Similar issues can be found in the implementation of whistleblower systems. Conceptually, whistleblower systems are considered one of the most effective mechanisms for uncovering fraud because they provide a safe channel for reporting violations (Herawati et al., 2023; N. W. A. Ningsih, 2020). However, the existence of the system alone does not automatically guarantee its effectiveness. Without a timely, professional, and adaptive response from auditors, incoming reports may not be followed up properly. Therefore, the effectiveness of a whistleblower system depends not only on the system itself, but also on the organizational environment and the capability of individual auditors to respond to reported information.
Same pattern also appears in professional skepticism. A critical and cautious mindset is generally believed to increase auditors’ sensitivity to fraud indicators (Akbar et al., 2022; Suriani et al., 2023). Nevertheless, empirical findings remain inconsistent, as several studies found no significant interaction between professional skepticism and fraud detection (Gunawan et al., 2022; Rahmawati et al., 2021). These inconsistent results suggest that professional skepticism alone may not be sufficient. Auditors also need additional capabilities that enable them to translate skeptical judgment into responsive and effective actions.
Based on these empirical gaps, this study argues that one important factor that has received limited attention is auditor agility. In today’s dynamic business environment, auditors are expected not only to possess technical competence, but also to be adaptive, flexible, and responsive to changing risks (Hussein Metwally & Ayoob Mohammed Hassan, 2023; Joshi, 2021). Auditor agility can be understood as a dynamic capability that enables auditors to utilize technology effectively, respond to information in real time, and adjust audit strategies according to the complexity of each case (DeRoche, 2022; PWC, 2020; Sasviranti et al., 2023). Existing research on auditor agility has primarily focused on its direct impact on audit performance, with only a few studies investigating its moderating function in the interaction of information systems, whistleblower mechanisms, professional skepticism, and fraud detection.
Accordingly, the novelty of this study lies in positioning auditor agility as a moderating variable from a dynamic capability perspective to explain the inconsistencies found in previous studies. This research not only examines the direct effects of information systems, whistleblower systems, and professional skepticism on fraud detection, but also explores how auditors’ adaptive capabilities may strengthen or weaken the effectiveness of organizational systems and individual characteristics in detecting fraud.
The importance of this study becomes even more relevant considering the large proportion of fraud losses that cannot be recovered and the increasingly sophisticated fraud schemes emerging in the digital era. As a result, the purpose of this study is to experimentally investigate the function of auditor agility in moderating information systems, whistleblower systems, and professional skepticism on fraud detection. As a result, the study is expected to make both theoretical and practical contributions to improving Indonesia's fraud detection procedures.
Based on the practical issues and theoretical gaps discussed above, this study specifically aims to examine the effects of information systems, whistleblower systems, and professional skepticism on fraud detection, as well as to analyze the moderating role of auditor agility in these relationships. In addition, this study seeks to explain the inconsistent findings in prior research by integrating the dynamic capability perspective into the context of fraud detection in the digital era. Ultimately, this study is expected to contribute theoretically by strengthening the conceptual model of fraud detection and practically by supporting the development of more effective audit strategies in Indonesia.
2. Literature Review and Hypotheses Development
2.1. Fraud Detection
Fraud detection has become an increasingly important issue in organizational governance due to the growing complexity of business processes and the rapid development of information technology. Fraud is generally understood as an intentional act carried out to obtain personal or group benefits at the expense of other parties. In the context of auditing, Fraud detection is the auditor's ability to identify indications, anomalies, or evidence of fraudulent activities during the audit process.
Various fraud theories have been developed to explain the factors that drive fraudulent behavior. The Fraud Triangle Theory, first proposed by Donald Cressey in 1953, is among the oldest and most well-known. It later served as the basis for a number of other theories about fraud, such as the Fraud Diamond Theory by David T. Wolfe and Dana R. Hermanson, 2004, the Fraud Pentagon Theory by Jonathan Marks, 2011, and the Fraud Hexagon Theory by Georgios L. Vousinas, 2019. More recently, the Fraud New Star Theory proposed by (Monteverde, 2021) expanded these perspectives by suggesting that fraud is influenced not only by pressure, opportunity, rationalization, and individual capability, but also by internal and external organizational factors as well as organizational social culture. According to this viewpoint, auditor conduct and the dynamics of the audit environment influence the efficacy of fraud detection in addition to organizational mechanisms.
In this study, information systems, whistleblower systems, and professional skepticism are positioned as the main factors influencing fraud detection. Meanwhile, the increasing complexity of fraud requires auditors to be more adaptive and responsive in dealing with changing risks. Therefore, auditor agility is considered a relevant factor that may strengthen the influence of these variables on fraud detection.
2.2. Dynamic Capabilities Theory
This study adopts Dynamic Capabilities Theory as the main theoretical foundation to explain how organizations and auditors manage resources and capabilities in responding to dynamic environments. According to (Teece et al., 1997), The ability of a company to integrate, develop, and reconfigure internal and external skills in order to successfully respond to changes in the environment is referred to as dynamic capabilities.
In the context of this study, information systems and whistleblower systems are viewed as organizational resources that support the fraud detection process. However, the existence of these systems alone may not automatically improve the effectiveness of fraud detection if auditors do not have the capability to utilize and adapt them to changing fraud risks. Professional skepticism is considered an ordinary capability embedded within the audit process, while auditor agility represents a dynamic capability that enables auditors to respond to change, adjust audit procedures, and integrate various sources of information in a flexible manner.
Therefore, Dynamic Capabilities Theory supports the argument that auditor agility may strengthen the effectiveness of information systems, whistleblower systems, and professional skepticism in improving fraud detection.
2.3. Information System and Fraud Detection
Information systems play an important role in supporting organizational transparency, internal control, and fraud detection processes. (Romney & Steinbart, 2020) explained, a system for gathering, processing, storing, and disseminating data to aid in decision-making is called an information system.
In the audit environment, information systems provide real-time, accurate, and traceable data that help auditors identify anomalies or unusual transactions more effectively. The implementation of integrated information systems, supported by strong internal controls, may reduce opportunities for fraud. The Committee of Sponsoring Organizations of the Treadway Commission (COSO) Internal Control Integrated Framework explains the control environment, risk assessment, control activities, information and communication, and monitoring activities are the five primary elements of an effective internal control system.
Previous studies have shown that information systems influence fraud detection. (Damayanti & Nugroho, 2024; Nurhasanah et al., 2019; Yusuf Ansori & Kuntadi, 2022), as well as (Najatu Silva & Aufa, 2023) found that information systems significantly affect fraud detection and fraud prevention. However, several studies reported inconsistent findings. (Choirunnisa, 2022; M. Ikhwan Mansyuri et al., 2023; Meiryani, et al., 2023; Polontalo et al., 2022) concluded that fraud detection is not much impacted by information systems.
These inconsistent findings suggest that the effectiveness of information systems may depend on the auditor’s ability to utilize the system and interpret the information provided. Auditors who are adaptive and responsive are more likely to integrate technology effectively into audit procedures, which may improve fraud detection.
H1: Information systems influence fraud detection.
2.4. Whistleblower System and Fraud Detection
A whistleblower system is a formal mechanism that allows individuals within an organization to report misconduct or suspected fraud in a safe and confidential manner. The system is built on the principles of trust, protection, and impartiality, ensuring that individuals feel secure when reporting information (COSO, 2013; International Organization for Standarization (ISO), 2021).
Whistleblower systems are considered important because fraud is often hidden and difficult to detect through routine audit procedures. Information provided by whistleblowers frequently serves as an early warning signal of irregularities within an organization (Hartono Putri & Trisnaningsih, 2022). Therefore, whistleblower systems are closely related to forensic auditing and investigative auditing practices.
Several studies have shown that whistleblower systems positively influence fraud detection. (Herawati et al., 2023; Kawistara Yasa & Haq, 2023; Piserah et al., 2022; Priyadi et al., 2022; Rahman, 2020) found that whistleblower systems improve the effectiveness of fraud detection. In addition, (Hamilah et al., 2022) and (Nwakeze et al., 2023) emphasized that stronger protection for whistleblowers enhances the effectiveness of whistleblowing systems.
However, the effectiveness of a whistleblower system also depends on the auditor’s ability to respond to and follow up on the reports received. Agile auditors are more likely to respond quickly and integrate whistleblower reports more effectively into audit procedures and fraud investigations.
H2: Whistleblower systems influence fraud detection.
2.5. Professional Skepticism and Fraud Detection
Professional skepticism describes the auditor’s critical mindset and questioning attitude when evaluating audit evidence. (Herawati et al., 2023) and (Kuntadi et al., 2022) reveals that auditors with a high level of professional skepticism are less likely to accept management’s explanations without conducting further examination and are more likely to be alert to signs of fraud.
According to (Suriani et al., 2023) and (Nurleni et al., 2024), professional skepticism plays an important role in fraud detection because fraudulent activities are usually concealed and intentionally designed to avoid detection. Auditors who maintain a skeptical attitude will more often seek additional evidence, identify inconsistencies in the information, and critically evaluate the reasonableness of transactions.
Numerous earlier studies have demonstrated that professional skepticism favorably affects fraud detection. (Akbar et al., 2022; Kuntadi et al., 2022; Nurleni et al., 2024; Pamela Citranagari, 2022; Pramuki & Agustine, 2023; Salsabila et al., 2023; Suriani et al., 2023; Surtikanti et al., 2023) discovered a strong correlation between professional skepticism and the auditor’s ability to detect fraud. However, other studies reported different findings. (Bagus Amlayasa & Riasning, 2022; Gunawan et al., 2022; Jumaili et al., 2023; Piserah et al., 2022; Rahmawati et al., 2021) discovered that fraud detection is not considerably impacted by expert skepticism. In addition, (Ardiansyah et al., 2024; N. Ningsih et al., 2022) suggested that factors such as auditor experience, time pressure, and task complexity may contribute to these inconsistent findings.
These differences indicate that professional skepticism alone may not be sufficient to improve fraud detection in dynamic and complex audit environments. Auditors also need adaptive capabilities and flexibility in applying skeptical judgment according to the audit conditions they face.
H3: Professional skepticism influences fraud detection.
2.6. Auditor Agility: A Direct Predictor and Moderating Variable in Fraud Detection
Auditor agility refers to the auditor’s ability to adapt quickly to change, respond effectively to risks, and adjust audit approaches according to the dynamics of the business environment. Agile auditors emphasize flexibility, collaboration, responsiveness, and continuous improvement throughout the audit process.
Rapid technological developments, increasingly complex fraud schemes, and changes in the business environment require auditors to possess adaptive capabilities beyond traditional audit procedures. Agile auditors are generally better equipped to manage large volumes of information, integrate data from multiple sources, and respond more quickly to fraud indicators.
Agile auditing is crucial for increasing audit effectiveness, according to earlier research. (Dobrowolski, 2021; Joshi, 2021; Koerniawati, 2021), as well as (Hussein Metwally & Ayoob Mohammed Hassan, 2023) explained that agile audit approaches improve responsiveness, communication, and audit efficiency. In this study, auditor agility is positioned both as a variable that directly influences fraud detection and as a moderating factor that improves the connection between information systems, whistleblower systems, professional skepticism, and fraud detection.
Auditor agility enables auditors to utilize information systems more effectively, respond to whistleblower reports more quickly, and apply professional skepticism more flexibly in complex audit situations. Thus, it is anticipated that auditor agility would improve the connection between these factors and fraud detection.
H4: Auditor agility influences fraud detection.
H5: Auditor agility strengthens the influence of information systems on fraud detection.
H6: Auditor agility strengthens the influence of whistleblower systems on fraud detection.
H7: Auditor agility strengthens the influence of professional skepticism on fraud detection.
According the theoretical and empirical review, this study proposes that information systems, whistleblower systems, and professional skepticism have direct effects on fraud detection. However, the effectiveness of these variables may vary depending on the auditor’s adaptive capabilities.
In this study, auditor agility is positioned as a moderating variable that enhances the impact of information systems, whistleblower systems, and professional skepticism on fraud detection in addition to being a variable that directly affects fraud detection. Auditors with a higher level of agility are expected to be better able to utilize technology, integrate information, and adjust audit procedures in response to the dynamic nature of fraud risks.
3. Method
3.1. Study Design
This research employed a quantitative approach using an explanatory research design to examine the causal relationships between information systems, whistleblower systems, professional skepticism, and fraud detection, auditor agility acts as both an independent and moderating variable. The quantitative approach was selected because the study relied on numerical data collected through questionnaires and analyzed using statistical techniques. In addition, the explanatory design was considered appropriate because the study aimed not only to describe phenomena, but also to explain Investigate causal linkages between variables and test the given hypotheses.
The study adopted a survey-based research design to obtain primary data directly from respondents. The population consisted of internal auditors in Indonesia, considering that internal auditors are directly involved in internal control, fraud prevention, and fraud investigation processes. Therefore, they were considered the most relevant respondents for examining factors influencing fraud detection. Purposive sampling was used to determine the sample, which included the following criteria:
(1) respondents must be internal auditors working in internal audit units in Indonesia, and
(2) respondents must have at least one year of audit experience.
3.2. Data Collection
Using the Slovin formula and a 10% margin of error, the minimum sample size required for this study was 92 respondents. Data were collected through an online questionnaire sent using Google Forms via WhatsApp groups inside the D'Indonesia IA Community, and other internal auditor networks outside the community. The questionnaire used a Likert scale to measure all research variables, with items rated on a five-point Likert scale (1 = strongly disagree to 5 = strongly agree).
This study examined five variables: information system, whistleblower system, professional skepticism, auditor agility, and fraud detection.
· Fraud detection adapted (Monteverde, 2021):Pressure, opportunity, Rasionalitation, capanility, Eksternal Factor, Internal Factor, Social/Organizational Culture.
· Information system was measured using indicators related to (COSO, 2013; Damayanti & Nugroho, 2024): system integration, data accuracy, information security, real-time reporting, accessibility.
· Whistleblower system was measured through (International Organization for Standarization (ISO), 2021; Nwakeze et al., 2023; Pamela Citranagari, 2022): accessibility, confidentiality, protection, responsiveness, and independence of the reporting mechanism.
· Professional skepticism was measured using indicators adapted (Suriani et al., 2023) such as critical thinking, questioning mind, scientific and strong evidence, openness in judgment, and due care.
· Auditor agility was measured through adapted (Dobrowolski, 2021): stakeholder focus, audit simplification, adaptive excellence, collaboration, process progress, motivation and sustainability, team independence, timely reporting, and continuous improvement.
3.3. Data Analysis
The data were analyzed with SmartPLS version 4.1.1.4, which used Partial Least Squares-Structural Equation Modeling (PLS-SEM). PLS-SEM was used for various methodological reasons. First, this study was predictive and explanatory in nature, aiming to examine and predict the relationships among variables as well as the moderating role of auditor agility. Second, the research model involved complex latent constructs and moderation effects, making PLS-SEM more suitable and flexible for estimating structural relationships. Third, PLS-SEM does not require multivariate normal data distribution, which is appropriate for questionnaire-based data using Likert scales.
To minimize the risk of common method bias (CMB), this study applied procedural remedies recommended by (Podsakoff et al, 2003), including ensuring respondent anonymity, maintaining confidentiality, and using neutral questionnaire wording. In addition, statistical tests for CMB were conducted using Harman’s Single Factor Test through Principal Component Analysis (PCA) and Full Collinearity Assessment using Variance Inflation Factor (VIF).
The analysis procedure consisted of descriptive statistics, measurement model evaluation (outer model), and structural model evaluation (inner model). The outer model evaluation included tests of convergent validity, discriminant validity, and construct reliability using outer loading values, Average Variance Extracted (AVE), Composite Reliability, and Cronbach’s Alpha. Indicators with loading values above 0.70 were considered acceptable, although loading values above 0.60 were still considered adequate for exploratory research. The inner model evaluation was conducted to test the proposed hypotheses using path coefficients, t-statistics, p-values, coefficient of determination (R²), and moderation effect analysis.
4. Results and Discussion
4.1. Description of the Research Subject
The minimum sample size required was 92 respondents, a figure based on the Slovin formula with a 10% margin of error. A total of 197 respondents were successfully recruited, which met the minimum sample requirement. The majority of respondents were internal audit practitioners working in the agribusiness sector (41.6%), followed by the manufacturing sector (16.8%). In terms of audit experience, the majority of respondents had more than 10 years of experience as internal auditors (53.3%), while 21.8% had between 6 and 10 years of experience. The demographic characteristics of the respondents are presented in Table 1.
Table 1: Demographic Data of Respondents
Business Sector | Work Experience (years) | % Sector | |||||
1 | >1 - 2 | 3 - 5 | 6 - 10 | >10 | Total | % | |
Information Technology |
|
|
|
| 3 | 3 | 1.5% |
Transportation |
|
|
|
| 3 | 3 | 1.5% |
Food And Beverage |
|
| 2 | 2 |
| 4 | 2.0% |
Logistic & Supply Chain |
| 1 | 1 |
| 4 | 6 | 3.0% |
Miscellaneous |
| 1 | 2 | 1 | 3 | 7 | 3.6% |
Government |
| 1 |
| 3 | 7 | 11 | 5.6% |
Mining & Energy |
|
| 2 | 5 | 7 | 14 | 7.1% |
Financial /Bank | 2 | 1 | 2 | 4 | 6 | 15 | 7.6% |
Services / Insurance |
| 1 | 7 | 3 | 8 | 19 | 9.6% |
Manufacturing | 2 | 1 | 3 | 7 | 20 | 33 | 16.8% |
Agribusiness | 2 | 2 | 16 | 18 | 44 | 82 | 41.6% |
Total | 6 | 8 | 35 | 43 | 105 | 197 | 100.0% |
% Work Experience | 3.0% | 4.1% | 17.8% | 21.8% | 53.3% | 100.0% |
|
4.2. Descriptive Statistics
Descriptive statistics were used to describe respondents’ perceptions of fraud detection (FD), information systems (IS), whistleblower systems (WS), professional skepticism (PS), and auditor agility (AA). Overall, all variables showed relatively high mean values, indicating positive respondent perceptions toward the effectiveness of fraud detection practices, information systems, whistleblower systems, professional skepticism, and auditor agility.
The fraud detection variable recorded mean values ranging from 3.980 to 4.665, while information systems ranged from 3.594 to 4.183, indicating moderate to high perceptions of system effectiveness. Whistleblower systems showed consistently high mean values between 4.203 and 4.406, suggesting strong respondent agreement regarding the effectiveness of reporting mechanisms. Professional skepticism demonstrated the highest mean values among all variables, ranging from 4.416 to 4.741, indicating that respondents perceived skeptical attitudes as highly important in the audit process. Similarly, auditor agility demonstrated high mean values ranging from 3.858 to 4.635, reflecting positive perceptions of auditors’ adaptability and responsiveness in managing fraud risks.
In general, the standard deviation values indicate acceptable variation in responses, while the skewness and kurtosis values suggest that the data distribution remained within acceptable limits for further analysis using PLS-SEM.
4.3. Model Testing
4.3.1. Common Method Bias (CMB) Test
Common method bias (CMB) was assessed using Harman’s Single Factor Test through Principal Component Analysis (PCA) and the Full Collinearity Test using Variance Inflation Factor (VIF) values in SmartPLS 4.1.1.4. Test Harman’s Single Factor showed that no single factor dominated the overall variance, with the first component explaining only 32.9% of the variance, which is below the 50% threshold (Harman, 1976; Podsakoff et al., 2003). In addition, all VIF values in the Full Collinearity Test were below the essential threshold of 3.3, showing that the data were free from significant common method bias (Kock, 2015).
4.3.2. Convergen Validity
Outer loading values and Average Variance Extracted (AVE) were used to assess convergent validity. The correlation between each indicator and its relevant construct was assessed using outer loading values. Valid indicators were those with loading levels greater than 0.70, while values above 0.60 were still regarded as acceptable. Indicators with loading values below 0.60 were excluded from further analysis.
Table 2: Outer Loading Convergen Validity Test
Indicator | FD | SI | WS | PS | AA | Description |
FD1 (Pressure) | 0.46 |
|
|
|
| Invalid |
FD2 (Opportunity) | 0.64 |
|
|
|
| Valid |
FD3 (Rasionalitation) | 0.43 |
