American Journal of Economic and Management Business p-ISSN: XXXX-XXXX e-ISSN: 2835-5199 Vol. 4 No. 8 Agustus 2025 1388 Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia Stephani Marvita Dwi Alindy Universitas Telkom, Indonesia Email: stephani.alindy@gmail.com Abstract Indonesia’s shift to the Coretax digital system heightens the urgency for tax consulting firms to streamline routine processes through Robotic Process Automation (RPA) and Artificial Intelligence (AI). The purpose of this study is to determine whether technology, technical, organizational, client, and environmental factors significantly influence the desire of Tax Consultant Firms to implement RPA and AI. The research method is a quantitative survey, and research data were obtained by distributing online questionnaires on a scale of 1 to 5 via social media, WhatsApp applications, or LinkedIn. Respondents in this study were full-time employees with various accounting positions who worked at tax consulting firms in Indonesia. The dependent variable of this study is the Desire of Consulting Companies to implement RPA and AI. Meanwhile, the independent variables are technology, technical, organization, clients, and environment. The stages of data analysis include validity testing, reliability testing, outer model analysis, inner model analysis, and hypothesis testing. The results depend on the questionnaire responses and the tests carried out. The findings highlight that adoption is driven less by technical sophistication and external support than by usability, leadership backing, workforce trust, credible privacy safeguards, and market rivalry. Practically, managers should prioritize user-centric design, change management, trust-building, and robust data governance to accelerate RPA/AI initiatives in tax services. Keywords: Implementation of RPA and AI, technology, technical, clients, and tax consultants. INTRODUCTION In Indonesia, the tax system has undergone significant changes, shifting from separate services to an integrated digital service called Coretax, which has been implemented since January 1, 2025. Coretax will meet various taxpayers’ needs, such as payments, reporting, and monitoring the status of Tax Bills and Confirmation Letters on Requests for Explanation from the tax office. This step demonstrates Indonesia's courage in utilizing advanced digital technology in the taxation sector. According to Wildan (2022), digitalization is a necessity in facing these tax developments, so tax consultants, who act as mediators between taxpayers and the government, must respond positively and adapt to these changes. This is necessary for tax consultants to remain relevant amidst the wave of change. According to Darusallam (2022), tax consulting firms must transform towards digital transformation to capitalize on the opportunities offered by digitalization. If tax authorities have already transitioned to digital, and tax consultants fail to keep pace, opportunities will be lost. With the rapid advancement of technology and information, tax consulting services will transform into a more diverse service, enabling services to be provided not only by traditional consultants but also by taxologists and robots. Taxologists are professionals skilled in utilizing technology to Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1389 optimize a company's tax functions. Therefore, these services can be provided by professionals or through applications (Wildan, 2022). According to Fiqram (2025), the current taxation process is a significant challenge for many companies, where every step, from data input to reporting, requires precision and considerable time. Furthermore, many companies still rely on manual methods to fulfill their tax obligations, which often leads to problems such as late reporting, data mismatches, or even administrative sanctions (Lahann et al., 2019). Tax professionals today must navigate digital technology and complex regulations that are constantly evolving (Abdul Rashid et al., 2024; Puspita, 2023). They are required to adopt new technologies while still performing routine tasks that dominate daily work time and meet management expectations. One of the key functions of tax is integrating and analyzing large amounts of data. Simply adding employees may not be an effective solution to this challenge (Vishnevsky & Chekina, 2018). This is where Robotic Process Automation (RPA) comes in, assisting with the execution of various tax functions. RPA allows employees to focus on tasks that require critical thinking, reducing routine operational costs, reducing errors, and increasing the speed and quality of work, ultimately reducing audit risk (ElectroNeek Robotics Inc., 2021). Tambunan (2023) argues that the work of tax consultants is more focused on complex tasks than on cognitive and repetitive tasks that machines can perform. AI can conduct more thorough tax risk analysis, such as taxpayer compliance, so humans can use it to facilitate their work, not replace it. With the support of technology, where physical presence is no longer crucial, a tax consultant can offer their consulting services to clients in different countries. This must be supported by the government so that local tax consultants can penetrate the international market (Darussalam, 2022). According to Teribery (2025), robots cannot make decisions on their own. They simply carry out instructions. While users can leverage RPA to collect data for AI, the two technologies are not truly interconnected because RPA operates on a workflow basis. At the same time, AI focuses on data analysis and forecasting. Therefore, it can be concluded that humans are still needed within companies for decision-making. Based on the explanation above, the problem that arose before the use of RPA and AI was that the data input and reporting process was still done manually, requiring precision and considerable time. This often led to issues such as late reporting, data mismatches, or even administrative sanctions. The use of RPA and AI is expected to simplify repetitive and routine tasks, such as entering tax data or reconciling work reports, thereby reducing the possibility of errors and ensuring regulatory compliance. Human resources can also be allocated to more complex and valuable activities, such as consulting or further analysis that require more profound insight. However, the challenges faced in implementing RPA and AI are not easy due to the constant changes in tax regulations. Furthermore, advances in AI technology also pose a threat to tax consultants who do not keep up with technological trends or still adhere to conservative approaches. Compliance-related tasks could be taken over by AI, potentially disrupting the Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1390 profession if it does not adapt. Therefore, technology is often viewed as a tool that will replace humans. Hypothesis H1: Ease of use of technology significantly influences the desire of Tax Consulting Firms to implement RPA and AI. According to Trinh et al. (2020), perceived ease of use refers to how consumers or individuals view the technology or system adopted by a company. This perception stems from their beliefs and evaluations of the time, cost, or effort required by the new technology or system. Various studies indicate that this technology is complex, requiring higher technical skills and greater implementation effort for adoption (Perdana et al., 2022). Some people may be hesitant about this technology because it is still new to them. Users may need a long time to learn and implement it. Therefore, a faster adoption rate depends on how easy the technology is to understand and use (Baiod & Husain, 2024). H2: Relative advantages in the use of technology significantly influence the desire of Tax Consulting Firms to implement RPA and AI. According to Xu and Zainal (2025), relative advantage is a criterion used to compare new technologies with traditional methods or techniques. The relative advantage of digital technology can be seen in whether the technology can reduce operational responsibilities, increase operational efficiency, and align with the values of its users. According to Oh et al. (2022) and Park et al. (2022), employees are more likely to accept changes towards digitalization if they perceive that these new technologies and processes are more efficient, effective, and provide greater benefits compared to existing ones. Furthermore, if employees perceive that digital technology can improve their performance, productivity, and competitiveness, this tends to increase their positive attitude towards the change. Relative advantage also influences how people perceive the usefulness of new technologies, which is a strong factor in the acceptance and adoption of these technologies (Park et al., 2022). Lutfi et al. (2023), Seshadrinathan and Chandra (2021), and Shetty and Panda (2023) suggest that the impact of relative advantage on technology adoption is an important factor that has a positive relationship with innovation adoption. H3: The compatibility of using RPA and AI applications with the tax system significantly influences the desire of Tax Consulting Companies to implement RPA and AI. Compatibility is a key determinant of an individual's intention to engage in a particular behavior (Nordhoff et al., 2021). According to Rogers (2003), as noted by Abubakari et al. (2024), based on the diffusion of innovation theory, compatibility, as perceived by individuals in each context, is an important determinant of digital technology adoption. This compatibility concept encompasses an individual's perception of whether a digital technology aligns with their needs, experiences, and values within a given context. According to Seethamraju and Hecimovic (2023), the lack of equipment compatibility currently hinders the adoption of AI technology, particularly as it is still in its early stages of development. H4: The maturity of RPA and AI device applications with the tax system significantly influences the desire of Tax Consulting Companies to implement RPA and AI. According to Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1391 Weritz et al. (2020), adapting to rapid and continuous changes in the digital and business environment and achieving digital maturity requires a gradual development of capabilities or the achievement of objectives from the initial stage to the desired end stage. This requires active implementation of the digital transformation process with application maturity and continuous adjustment (Kane et al., 2017, in Ladu et al., 2024). Pre-defined measures and associated criteria describe the essential elements of an application's maturity, including key success factors, capabilities, resources, and areas of action (Ladu et al., 2024). H5: Top management commitment in an organization significantly influences the desire of Tax Consulting Firms to implement RPA and AI. According to Tarigan et al. (2020), top-level management support is crucial for establishing company goals, which is reflected in the decisions and actions implemented. Management support or commitment is the extent to which company management invests in technological innovation (Baiod & Husain, 2024). This support is one of the best indicators of an organization's adoption of new technology (Lutfi et al., 2023; Shetty & Panda, 2023). Top-level management support encompasses the extent to which management supports the team as a whole and helps resolve issues within the given timeframe. This term refers to the extent to which management supports the adoption of new technology (Akel & Ibrahim, 2020). H6: Employee trust in technology within an organization significantly influences the desire of Tax Consulting Firms to implement RPA and AI. According to Seethamraju and Hecimovic (2023), employee distrust of AI devices is another factor hindering technology adoption. AI technology can increase employee authority by providing more control over their tasks and work methods. This sense of empowerment increases employee job satisfaction and motivation (Lin, 2023). By minimizing the need for constant supervision and creating a more autonomous work environment, AI fosters trust and a sense of responsibility among employees (Adityaksa & Suyoso, 2025). However, technological limitations significantly contribute to employee distrust of new systems. This increases user stress and resistance to change, leading to a reluctance to learn how to use new technological solutions while clinging to existing ones (Valtonen & Holopainen, 2025). H7: Client concerns regarding data privacy significantly influence the desire of Tax Consulting Firms to implement RPA and AI. Privacy in AI means protecting confidential information, obtaining permission to access data, ensuring AI models do not leak protected data, using models ethically, and complying with global laws that not only recommend but also require maintaining privacy (Ammanath, 2022, p. 128). According to Kokina et al. (2025), companies must maintain the confidentiality of client data and build trust with them. Companies must be transparent about data use and demonstrate good governance to convince clients to share more data. According to Leroux and Pupion (2022), client concerns about privacy can be significantly reduced by demonstrating compliance with regulations, such as the General Data Protection Regulation (GDPR). H8: Client readiness regarding the use of technology will significantly influence the desire of Tax Consulting Firms to implement RPA and AI. According to Meuter et al. (2005) in Shim et al. (2020), consumer/client readiness includes role clarity, motivation, and capability. Role clarity Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1392 here is defined as the consumer's knowledge and understanding of their role. Motivation means having the intention to achieve something. Meanwhile, capability is self-confidence in mastering the skills needed to adopt new technology and complete a task. This is also reflected in the field of auditing. According to Pedrosa et al. (2020), client readiness, such as increased information technology connectivity and better data accessibility, enables firms to increase efficiency and accuracy in the audit process, but also encourages auditors to offer new assurance services to address risks arising from the adoption of new technologies. The use of AI by the client generally depends on the client's needs and how much value they place on AI, given that clients prefer to interact directly with consultants rather than using tools alone (Kokina et al., 2025). H9: Competitive pressure from the environment (competitors) significantly influences the desire of Tax Consulting Firms to implement RPA and AI. Intense competition has long been recognized as a significant factor driving the adoption of new technologies (Lutfi, 2022; Sharma et al., 2023). This competition reflects the pressure an organization feels from competitors in its industry (Lutfi, 2022). When many competitors have implemented new technologies, companies also need to follow suit by adopting similar technologies to advance business processes and remain competitive (Siew et al., 2020). The implementation of this new technology will significantly increase a company's competitiveness by improving business processes as well as the quality and efficiency of its business practices (Baiod & Husain, 2024). According to Shahadat et al. (2023), fierce competition in an industry can force managers to adopt technologies widely used in that sector. Under competitive pressure, companies often adopt digital technologies to stay ahead. H10: Vendor support significantly influences the desire of Tax Consulting Firms to implement RPA and AI. With increasing industrial competition, globalization, and technological disruption, the importance of building strong relationships with vendors has increased significantly (Oyedokun et al., 2024; Segun-Falade et al., 2024). In this digital age, relationships between companies and vendors are more than just transactions, but rather long-term partnerships that require continuous collaboration and optimization (George et al., 2024). This shift is driven by trends such as globalization, which has expanded the scope and diversity of vendor networks, and digital transformation, which enables the integration of new technologies into vendor management systems (Okeke et al., 2024). According to Baiod and Husain (2024), with limited IT expertise within a company, companies may rely on support from technology providers. Organizations are more willing to take risks and try new technologies if they feel adequately supported by their suppliers or third parties. METHOD The research method was a quantitative survey, and research data were obtained by distributing online questionnaires on a scale of 1 to 5 via social media, WhatsApp applications, or LinkedIn. Respondents in this study were full-time employees with various accounting positions who worked at tax consulting firms in Indonesia. This research was implemented with a level of accuracy (α) of 5% and a confidence level of 95%, which resulted in a Z value of 1.96. The probability of the questionnaire being Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1393 correct/accepted was 0.5, and the probability of it being incorrect or rejected was also 0.5. Therefore, by using the Bernoulli formula, the minimum sample required was 100 respondents, consisting of full-time employees holding various accounting positions working in tax consulting firms in Indonesia. In this study, data processing was carried out using path analysis with the Partial Least Squares (PLS) method. The data analysis techniques used were the Measurement Model (Outer Model), Inner Model, and Hypothesis Testing. RESULT AND DISCUSSION The correlation method was calculated using SPSS version 31. Validity testing was used to assess the validity of a questionnaire. Reliability testing was used to measure the consistency of the research variables. This test was conducted early in the research as a pre-test, distributing questionnaires to 50 respondents who met the eligibility criteria. The method used to assess the validity of the questionnaire is the Product-Moment correlation or the Bivariate Pearson correlation. By using the r-value distribution table of the Product-Moment Sig. 5%, then with a sample of 50 responses, the r-table value is 0.279. Thus, the statement is valid if the calculated r- value is > 0.279. The result is that all amounts are greater than 0.279, so it can be said that all statements in the questionnaire are valid. The Cronbach's Alpha value, either in total or if one of the items is removed, remains the same. Greater than 0.6 (the result is 0,966), for all variables for a sample size of 50 respondents, so it can be said that all statements in the questionnaire are reliable. Testing on Population Outer Model There are two measurement models in the outer model, namely the reflective measurement model and the formative measurement model. The Reflective Measurement Model consists of: 1) Test Convergent Validity The convergent validity test aims to assess the validity of each correlation between indicators and latent variables. A correlation is considered to meet convergent validity if it has a factor loading value of 0.70 (Ghozali, 2021, p. 68). several variable indicators have values below 0.70, indicating that these variables are more correlated with other constructs than they should be. These invalid indicators can lead to inaccurate research conclusions. Therefore, it is necessary to identify outliers for several variables that do not meet the validity requirements. This is achieved by eliminating one or more indicators that do not have a strong relationship with the latent variable. all indicators for each variable have met the convergent validity requirements with values for each indicator greater than 0.70. Loading factors greater than 0.7 indicate that the indicators are valid for measuring the intended construction and demonstrate good discriminant validity. 2) Test Discriminant Validity Discriminant validity can be measured using three methods, namely Average Variance Extracted (AVE), Cross Loading, and Heterotrait-monotrait (HTMT). Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1394 a. Average Variance Extracted (AVE) The Fornell-Larcker criterion aims to compare the square root of the Average Variance Extracted (AVE) of each construct with the correlation between other constructs in the model (Henseler et al., 2015:127 in Sönmezay, 2023). If the square root of the AVE of each construct is greater than the correlation value between that construct and other constructs in the model, then the model is said to have good discriminant validity (Ghozali, 2021, p. 69). Table 1. Average Variance Extracted (AVE) Value Variables AVE - Count AVE - Criteria Information Technology - Easy to Use 0.602 0.500 valid Technology - Relative Advantages 0.660 0.500 valid Technical - Compatibility 0.682 0.500 valid Technical - Application Maturity 0.730 0.500 valid Organization - Top Management Commitment 0.660 0.500 Valid Organization - Employee Trust 0.758 0.500 valid Client - Data Privacy 0.683 0.500 valid Client - Client Readiness 0.766 0.500 valid Environment - Competitive Pressure 0.719 0.500 valid Environment - Vendor Support 0.595 0.500 valid Consulting Firms' Desire to Implement RPA and AI 0.839 0.500 valid Based on Table 1 above, the AVE value of all variables is greater than 0.50, indicating that the indicators used to measure these variables have good convergent validity and are acceptable for further analysis. a. Cross Loading The cross-loading value of each construct was evaluated to ensure that the correlation between the construct and the measurement items was greater than that of the other constructs. The expected cross-loading value is greater than 0.7 (Ghozali, 2021, p. 68). the cross-loading value of the indicators has exceeded 0.7, so that each indicator is considered valid and shows that the constructs are conceptually different and can be measured separately. b. Heterotrait-Monotrait (HTMT) The Heterotrait-Monotrait ratio (HTMT) is one of the criteria for testing discriminant validity in variance-based structural equation modeling, particularly those using SmartPLS. The generally accepted HTMT value is less than 0.90. A value higher than 0.90 indicates a discriminant Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1395 validity problem, meaning different constructs may not truly differ from each other in the model (Roemer & Henseler, 2021). Table 2. HTMT value ( Heterotrait-Monotrait ratio) Variable Kei KliKes KliPri LingDung LingTek OrKep OrKom TekKem TekKeu TekKom TekMud Kei KliKes 0.353 KliPri 0.775 0.499 LingDung 0.521 0.424 0.702 LingTek 0.660 0.449 0.641 0.780 OrKep 0.674 0.805 0.694 0.654 0.418 OrKom 0.733 0.474 0.706 0.796 0.783 0.669 TekKem 0.620 0.111 0.538 0.531 0.472 0.297 0.508 TekKeu 0.469 0.079 0.276 0.179 0.307 0.132 0.259 0.611 TekKom 0.668 0.304 0.758 0.506 0.449 0.590 0.499 0.757 0.349 TekMud 0.836 0.304 0.778 0.545 0.547 0.551 0.695 0.788 0.503 0.849 Based on Table 2 above, the HTMT value for each indicator is less than 0.9, so discriminant validity is considered met. This indicates that the different constructs are indeed statistically different. 3) Test Internal Consistency Reliability Methods for measuring Internal Consistency: Reliability tests using Cronbach's Alpha and Composite Reliability. a. Cronbach's Alpha According to Ghozali (2021), Cronbach's alpha measures explicitly internal consistency, namely the extent to which items in a measurement instrument correlate with each other and measure the same concept. Interpretation of Cronbach's Alpha Values: a. > 0.7: Considered good and indicates high internal consistency. b. 0.6 - 0.7: Acceptable, but there may still be room for improvement. c. < 0.6: Indicates low reliability and the instrument may need to be improved or reconsidered. above, six (6) variables are good, and three (3) variables are still acceptable. Meanwhile, two (2) variables have low reliability, namely the Technology-Relative Advantage variable and the Technical-Compatibility variable. b. Composite Reliability According to Ghozali (2021), Composite reliability (CR) is a method used in scientific research to assess the consistency of measuring a construct or variable through its indicators. In the context of Structural Equation Modeling (SEM), CR is considered better in estimating reliability because it takes into account the loading factor of each indicator. A high CR value indicates that the indicators are reliable in consistently measuring the same variable. Generally, an acceptable CR value is greater than 0.7, but a value greater than 0.6 is still considered adequate. Table 3. Composite reliability value Variables CR - Count CR - Criteria Reliability Technology - Easy to Use 0.858 0.700 Good Technology - Relative Advantages 0.794 0.700 Good Technical - Compatibility 0.809 0.700 Good Technical - Application Maturity 0.843 0.700 Good Organization - Top Management Commitment 0.853 0.700 Good Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1396 Variables CR - Count CR - Criteria Reliability Organization - Employee Trust 0.863 0.700 Good Client - Data Privacy 0.865 0.700 Good Client - Client Readiness 0.908 0.700 Good Environment - Competitive Pressure 0.835 0.700 Good Environment - Vendor Support 0.854 0.700 Good Consulting Firms' Desire to Implement RPA and AI 0.912 0.700 Good Based on Table 3 above, the Composite reliability value for each indicator is greater than 0.7, indicating that the variable is acceptable and relatively consistent, and is likely to be good and stronger because the CR value is mostly more than 0.8. Only one (1) variable, namely the Technology-Relative Advantage variable, has a CR below 0.8 but remains above 0.7. The Formative Measurement Model consists of: a. Weight Value If the weight value is statistically significant, it indicates that the indicator has a significant impact on the construct. Therefore, the estimated value in the formative measurement model must be significant (P < 0.05). This level of significance is assessed using a bootstrapping procedure (Ghozali, 2014, in Jesica, 2020). the p-value for each indicator is less than 0.05, indicating that the indicator significantly contributes to the measurement of its latent variable. This means the indicator can represent or reflect the latent variable it represents. b. Multicollinearity Multicollinearity in the outer model occurs when the indicators of the latent variable are highly correlated. To determine whether a formative indicator experiences multicollinearity, a Variance Inflation Factor (VIF) value of <10 indicates that the indicator does not experience multicollinearity ( Ghozali, 2014, in Jesica, 2020 ). the Variance Inflation Factor (VIF) value for each indicator is less than 10, indicating that there is no multicollinearity problem between indicators in a latent variable. In other words, the indicator variables used to measure a latent variable are not highly correlated with each other. Inner Model The Inner Model is a structural model that describes the causal relationship between latent variables formulated based on the underlying theory. (Ghozali, 2014, in Jesica, 2020). The following are the criteria for assessing the inner model: 1) Model Fit Test SmartPLS has several different model fit criteria (Ringle et al., 2024), and a model is considered fit if it meets most of these criteria, not all. Table 4. Model Fit Analysis Parameter Criteria Results Information SRMR Less than 0.10 0.08 fit d_ULS > 0.05 3 fit d_G > 0.05 1 fit Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1397 X² X² statistic < x² table 717,287 > 41,337 not fit NFI approaching the value 1 0.553 fit GOF 0.1 (small GOF), 0.25 (moderate GOF), 0.36 (strong GOF) 0.653 fit (strong) Q² > 0 0.503 fit The chi-square test consists of 29 indicators, so the dk value used in the chi-square table is dk = 28 with a significance level of 5%. Based on Table 4 above, six criteria have been successfully met, while only one criterion is not met, namely the chi-square test. The chi-square test only shows whether there is a significant relationship between variables. However, it does not provide information about the strength or direction of the relationship, so the chi-square criteria are often inappropriate. Because most of the model fit criteria are met, the model is appropriate to the data 2) Coefficient of Determination Test (R-squared) This method is used to determine the effect of independent variables on the dependent variable. A high R-squared value (close to 1) indicates that the model can explain most of the variation in the data, while a low value (close to 0) indicates that the model is less able to explain that variation. An R² value of 0.75 is considered good, 0.50 moderate, and 0.25 weak (Ghozali, 2021). Table 5. R-Square Value R-square R-square adjusted Consulting Firms' Desire to Implement RPA and AI 0.610 0.568 Based on Table 5 above, the R Square value for the simultaneous influence of the independent variables on the dependent variable is 0.610 with an adjusted R Square value of 0.568. Therefore, it can be explained that all exogenous constructs (independent variables) simultaneously influence the dependent variable by 56.8%, which is moderate. 3) Effect Size (F-squared or F²) F-squared is used to assess model quality and provide more specific information about the strength of the influence of a particular independent variable, which helps understand its importance in the model. Recommended effect sizes are 0.02, 0.15, and 0.35, representing small, moderate, and significant effects of the independent latent variable at the structural level (Ghozali, 2021). the Client variable - Client Readiness, Technical - Application Maturity, Technology - Relative Advantage, and Technical - Compatibility have weak or even insignificant influence on RPA and AI Adoption. Hypothesis Testing Hypothesis testing is used to explain the direction of the relationship between dependent and independent variables. Hypothesis testing is performed by examining probability values and Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1398 t-statistics. For probability values and p-values with a 5% alpha, they are <0.05. According to Ghozali and Latan (2020:42), the interpretation at a 5% significance level is as follows: 1) P-value < 0.05: H₀ is rejected or Ha is accepted, meaning the dependent variable has a significant effect on the independent variable. 2) P-value ≥ 0.05: H₀ is accepted or Ha is rejected; the dependent variable does not have a significant effect on the independent variable. This means that the p-value (probability) indicates the likelihood of generalizing the data to the population, with a 5% chance of error and a 95% confidence level in decision-making. In this study, the hypothesis test uses the path coefficient, which helps indicate the direction of the relationship between the research variables. Path coefficients can be used to determine whether a hypothesis is positive or negative (Ghozali, 2021). Table 6. Path Coefficient Value Hypothesis Variables T statistics (|O/STDEV|) P value - Calculate P value - Criteria Information H1 Technology - Easy to Use → Implementation of RPA and AI 2.318 0.010 P < 0.05 significant H2 Technology - Relative Advantages → Implementation of RPA and AI 1.469 0.071 P < 0.05 insignificant H3 Technical - Compatibility → Implementation of RPA and AI 0.110 0.456 P < 0.05 insignificant H4 Technical - Application Maturity → Implementation of RPA and AI 1.132 0.129 P < 0.05 insignificant H5 Organization - Top Management Commitment → Implementation of RPA and AI 2.085 0.019 P < 0.05 significant H6 Organization - Employee Trust → Implementation of RPA and AI 3.174 0.001 P < 0.05 significant H7 Client - Data Privacy → Implementation of RPA and AI 1.895 0.029 P < 0.05 significant H8 Client - Client Readiness → Implementation of RPA and AI 1.268 0.102 P < 0.05 insignificant H9 Environment - Competitive Pressure → Implementation of RPA and AI 1.908 0.028 P < 0.05 significant H10 Environment - Vendor Support → Implementation of RPA and AI 1.313 0.095 P < 0.05 insignificant five test results show a significant positive influence, namely the variables Client - Data Privacy, Environment - Competitive Pressure, Organization - Employee Trust, Organization - Top Management Commitment, and Technology - Ease of Use. The remaining variables, Client - Client Readiness, Environment - Vendor Support, Technical - Application Maturity, Technical - Compatibility, and Technology - Relative Advantage, show insignificant test results with a p-value >0.05. Based on the results of the hypothesis testing above, it can be concluded that: Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1399 1) Hypothesis 1 (H1) The p-value of 0.010 is lower than the 0.05 significance level, so Ho is rejected and Ha is accepted. Therefore, it can be concluded that the ease of use of technology significantly influences the willingness of Tax Consulting Firms to implement RPA and AI. RPA and AI technologies are complex and require higher technical skills (Perdana et al., 2022), so ease of use is an important consideration in implementing RPA and AI. This is in line with research by Trinh et. al. (2020), perceived ease of use refers to how consumers or individuals view the technology or system adopted by a company. This perception stems from their beliefs and evaluations of the time, cost, or effort required by the new technology or system. Therefore, faster adoption rates depend on how easy the technology is to understand and use (Baiod & Husain, 2024). 2) Hypothesis 2 (H2) The p-value of 0.071 is greater than the 0.05 significance level, so Ho is accepted and Ha is rejected. Therefore, it can be concluded that the relative advantage of technology use does not significantly influence the willingness of tax consulting firms to implement RPA and AI. Relative advantage is a criterion used to compare new technologies with traditional methods or techniques (Xu & Zainal, 2025). According to respondents, the advantages of using RPA and AI have not entirely influenced their implementation because traditional techniques are still considered more relevant, especially for most respondents who are millennials or generation Y, aged 29-44 years. According to Asni et. al. (2025), the Millennial Generation developed in the early digital era when available technology was still fundamental and limited, so traditional techniques are still considered more relevant in tax work. This is also in line with research by Park et. al. (2022), that relative advantage also influences how people view the usefulness of new technologies, which is a strong factor in the acceptance and adoption of these technologies. 3) Hypothesis 3 (H3) The p-value of 0.456 is greater than the significance level of 0.05, so Ho is accepted and Ha is rejected. Therefore, it can be concluded that the compatibility of RPA and AI applications with the tax system does not significantly influence the desire of Tax Consulting Firms to implement RPA and AI. Tax Consulting Firms believe that RPA and AI technology cannot yet be fully implemented because these technologies are still in the early stages of development. This is in line with the opinion of Seethamraju and Hecimovic (2023), who state that the lack of equipment compatibility currently hinders the adoption of AI technology, mainly because it is still in the early stages of development. 4) Hypothesis 4 (H4) The p-value of 0.129 is greater than the significance level of 0.05, so Ho is accepted and Ha is rejected. Therefore, it can be concluded that the maturity of RPA and AI device applications with the tax system does not significantly influence the desire of Tax Consulting Firms to implement RPA and AI. In line with the fourth hypothesis above, Tax Consulting Firms assume that RPA and AI technology cannot be fully implemented because these Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1400 technologies are still in the early stages of development. This is in line with the opinion of Weritz et. al. (2020), to adapt to rapid and continuous changes in the digital and business environment and to achieve a level of digital maturity, gradual development in capabilities or achievement of goals is required from the beginning to the desired final stage. 5) Hypothesis 5 (H5) The p-value of 0.019 is lower than the 0.05 significance level, so Ho is rejected and Ha is accepted. Therefore, it can be concluded that top management commitment within an organization significantly influences a Tax Consulting Firm's desire to implement RPA and AI. The success or failure of RPA and AI implementation requires the support of top management capable of making decisions about implementation steps. This also aligns with research conducted by Tarigan et al. (2020). Management support or commitment is the extent to which a company's management invests in technological innovation (Baiod & Husain, 2024). This support is one of the best indicators of an organization's adoption of new technology (Lutfi et al., 2023; Shetty & Panda, 2023). Top-level support encompasses the extent to which management supports the team as a whole and helps resolve issues within the given timeframe. This term refers to the level of support management provides when adopting new technology (Akel & Ibrahim, 2020). 6) Hypothesis 6 (H6) The p-value of 0.001 is lower than the 0.05 significance level, so Ho is rejected and Ha is accepted. Therefore, it can be concluded that employee trust in technology within an organization significantly influences the desire of Tax Consulting Firms to implement RPA and AI. This aligns with research conducted by Seethamraju and Hecimovic (2023), which found that employee distrust of AI devices is another factor hindering technology adoption. AI technology can increase employee authority by providing more control over their tasks and work methods. This sense of empowerment increases employee job satisfaction and motivation (Lin, 2023). By minimizing the need for constant supervision and creating a more autonomous work environment, AI fosters trust and a sense of responsibility among employees (Adityaksa & Suyoso, 2025). 7) Hypothesis 7 (H7) The p-value of 0.029 is lower than the 0.05 significance level, so Ho is rejected and Ha is accepted. Therefore, it can be concluded that client concerns regarding data privacy significantly influence tax consulting firms' willingness to implement RPA and AI. Data privacy is crucial, especially for tax consulting firms that hold important data from various companies. This aligns with research conducted by Kokina et al. (2025), which states that companies must maintain the confidentiality of client data and build trust with them. Companies must be transparent about data use and demonstrate good governance to convince clients to share more data. According to Leroux and Pupion (2022), client privacy concerns can be significantly reduced by demonstrating compliance with regulations, such as the General Data Protection Regulation (GDPR). 8) Hypothesis 8 (H8) Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1401 The p-value of 0.102 is greater than the 0.05 significance level, so Ho is accepted and Ha is rejected. Therefore, it can be concluded that client readiness for technology adoption will not significantly influence tax consulting firms' willingness to implement RPA and AI. According to Meuter et al. (2005) in Shim et al. (2020), consumer/client readiness includes role clarity, motivation, and ability. The use of AI generally depends on the client's needs and the value they place on AI, considering that clients prefer to interact directly with consultants rather than using tools alone (Kokina et al., 2025), so clients think that direct consultations are still more efficient than using RPA and AI technology. 9) Hypothesis 9 (H9) The p-value of 0.028 is lower than the 0.05 significance level, so Ho is rejected and Ha is accepted. Therefore, it can be concluded that competitive pressure from the environment (competitors) significantly influences the desire of tax consulting firms to implement RPA and AI. Intense competition has long been recognized as a significant factor driving the adoption of new technologies (Lutfi, 2022; Sharma et al., 2023). This competition reflects the pressure an organization feels from competitors in its industry. (Lutfi, 2022). When many competitors have implemented new technologies, companies also need to follow suit by adopting similar technologies to advance business processes and remain competitive (Siew et al., 2020). The implementation of this new technology will significantly increase a company's competitiveness by improving business processes as well as the quality and efficiency of its business practices ( Baiod & Husain, 2024). According to Shahadat et al. ( 2023 ), Fierce competition in an industry can force managers to adopt technologies widely used in that sector. Under competitive pressure, companies often adopt digital technologies to stay ahead. 10) Hypothesis 10 (H10) The p-value of 0.095 is greater than the 0.05 significance level, so Ho is accepted and Ha is rejected. Therefore, it can be concluded that vendor support does not significantly influence the desire of Tax Consulting Firms to implement RPA and AI. This is in line with research by Ifinedo (2011) in Baiod and Husain (2024), which found that external support from technology vendors had no significant relationship with the acceptance of internet technology or e-business. One reason is that most respondents did not value the importance of support from suppliers for emerging technologies in efforts to increase their adoption, or they did not have a favorable view of the vendors of these new technologies in their context. Respondents may have felt that support from vendors was inadequate and that there was still much room for improvement. In general, the technologies analyzed were quite new, and support from vendors could significantly influence adoption. CONCLUSION This research successfully met its objectives by identifying key factors influencing the desire of Tax Consulting Firms to implement RPA and AI. It found that ease of use of technology, top management commitment, employee trust in technology, client concerns about data privacy, and Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1402 competitive pressure significantly impact adoption intentions. In contrast, relative advantages of technology, compatibility with the tax system, application maturity, client readiness, and vendor support did not significantly affect this desire. Future research could explore deeper into the reasons behind the insignificant impact of these latter factors and investigate how evolving technological and regulatory environments may alter these dynamics over time. REFERENCES Abdul Rashid, S. F., Sanusi, S., & Abu Hassan, N. S. (2024). Digital Transformation: Confronting Governance, Sustainability, and Taxation Challenges in an Evolving Digital Landscape. In Corporate Governance and Sustainability: Navigating Malaysia’s Business Landscape (hal. 125–144). Springer. Abubakari, M. S., Zakaria, G. A. N., dan Musa, J. (2024). Perceived compatibility and students’ intention to adopt digital technologies in Islamic education institutions. Cogent Education, 11(1), 2430869. Adityaksa, R., dan Suyoso, A. L. A. (2025). The Impact of AI Adoption on Job Engagement and Employee Trust. Golden Ratio of Human Resource Management, 5(1), 133-140. Akel, S. A. A., dan Ibrahim, M. (2020). The Factors Affecting E-Filing Adoption Among Jordanian Firms: The Moderating Role of Trust. Palarch’s J. Archaeol. Egypt/Egyptol, 17, 14-31. Ammanath, B. (2022). Trustworthy AI: a business guide for navigating trust and ethics in AI. New Jersey: John Wiley & Sons. Asni, A., Chairunnisa, D., Salsabila, S., Putri, D. M., Alya, S., dan Ryansyah, M. R. (2025). Analisis GAP Generasi:: Generasi Milenial dan Generasi Z di Lingkungan Kampus. Indonesian Research Journal on Education, 5(1), 374-380. Cakrawala. (2023). Pengertian Teknologi: Jenis-Jenis dan Manfaatnya Bagi Kehidupan Manusia. [online]. https://www.cakrawala.ac.id/berita/teknologi- adalah#:~:text=Secara%20umum%2C%20pengertian%20teknologi%20adalah,berbagai%20perm asalahan%20dalam%20kehidupan%20manusia. [22 April 2025]. Darussalam. (2022). Webinar PKE 2022: Penerapan Ekonomi Digital: Penguatan dan Peran Konsultan Pajak dalam Praktik. [online]. https://www.youtube.com/watch?v=Qf3PwfBOdGc [22 April 2025]. ElectroNeek Robotics Inc. (2021). RPA in Tax Processes: What Can Be Automated?. [online]. https://electroneek.com/blog/rpa-in-tax-processes-10-use-cases/ [22 April 2025]. Fiqram, D. (2025). Otomatisasi Pajak di Era Digital dengan Coretax Automation. [online]. https://idstar.co.id/otomatisasi-pajak-di-era-digital-dengan-coretax-automation/ [22 April 2025]. George O. O., Dosumu, R. E., dan Makata, C.O. (2024). Strategic Vendor Relationship Management: A Conceptual Model for Building Sustainable Partnerships in Competitive Marketing Ecosystems. Journal of Frontiers in Multidisciplinary Research, 5(1), 112-118. Ghozali, I. (2021). Aplikasi Analisis Multivariate Dengan Program IBM SPSS 26 Edisi 10. Semarang: Badan Penerbit Universitas Diponegoro. Jesica, J. (2020). Analisis Pengaruh Persepsi Kemudahan dan Kemanfaatan Terhadap Perilaku Konsumen Untuk Minat Menggunakan Dompet Digital (Studi Pada Konsumen Starbucks Cabang Gambir). [online]. http://repository.stei.ac.id/2373/ [22 April 2025]. Kokina, J., Blanchette, S., Davenport, T. H., & Pachamanova, D. (2025). Challenges and opportunities for artificial intelligence in auditing: Evidence from the field. International Journal of Accounting Information Systems, 56, 100734. Lahann, J., Scheid, M., & Fettke, P. (2019). Utilizing machine learning techniques to reveal vat compliance Factors that Influence Companies to Implement the Use of RPA and AI in Tax Consulting Service Companies in Indonesia 1403 violations in accounting data. 2019 IEEE 21st conference on business informatics (CBI), 1, 1–10. Leroux, E., & Pupion, P. C. (2022). Smart territories and IoT adoption by local authorities: A question of trust, efficiency, and relationship with the citizen-user-taxpayer. Technological Forecasting and Social Change, 174, 121195. Puspita, T. (2023). Tax Planning in Transition: Evaluating the Impact of Evolving Legislation on Financial Strategies. Advances in Taxation Research, 1(3), 136–146. Vishnevsky, V. P., & Chekina, V. D. (2018). Robot vs. tax inspector or how the fourth industrial revolution will change the tax system: a review of problems and solutions. Journal of Tax Reform. 2018. Т. 4.№ 1, 4(1), 6–26. Copyright holders: Stephani Marvita Dwi Alindy (2025) First publication right: AJEMB – American Journal of Economic and Management Business