The technological revolution poses a constant challenge for accountants in terms of its adaptation. We aimed to empirically examine the factors influencing the use of AI in the accounting profession in Poland.
We based our study on the results of a survey conducted in 2021 among 231 corporate managers of small, medium, large and multinational corporations. We used the Paper & Pen Personal Interview method and surveyed MBA classes. We employed a logistic regression model with the moderating influences of size and organizational profile on the data collected from the survey.
The more organizations used “remote work” before the pandemic, the more willing they are now to adopt AI technologies in accounting. Moreover, the findings show that organizations, to a smaller extent, plan to change their business model after the pandemic and are more willing to adopt AI technologies. Organizational culture and values are not significant factors, which is contrary to what we expected.
We empirically investigated the impact of remote work, both before and during the COVID-19 pandemic, and business model changes regarding the AI use in the accounting profession. We consider these relationships to be innovative, thereby contributing to the existing literature. We used logistic regression considering the moderating influences of size and organizational profile.
Introduction
Information processing and communication technologies have profound implications for accounting practice (Hunton, 2022). The accountant's role is transforming. More and more often, AI is taking over traditional responsibilities, e.g. posting accounting records. Today, accountants are expected to contribute to strategic decision-making processes and solve key management issues within the company. The accountant profile requires new skills and competencies, i.e. reorganizing economic processes, creating and interacting with new business models and e-models, and conducting consulting and strategic management. Nimigean (2021) identified the traditional fundamental characteristics of the accounting profession, such as knowledge, professionalism, reliability and honesty, which were and remain invaluable in today's world. However, the technological revolution has increased the importance of other attributes and skills, such as group work, adaptation to new technologies, problem-solving attitude and time management. The digitalization of accounting processes is not a new concept, and many companies have used it for a long time. However, the pandemic forced other companies, more technology-opportunistic, to implement AI technologies or accelerate the implementation pace.
The technological revolution is a constant challenge for accountants in terms of adaptation. It also presents the threat of redundancy in accounting-related professions, which may eventually lead to job losses. These and similar views are partially justified. However, this is not a complete picture. Gulin, Hladika, and Valenta (2019) note that a substantial part of the accounting community still upholds a traditional perspective and perceives their profession as based on repetitive tasks and following routine schemes. They postulate a change of this attitude and recommend that accountants adopt new roles as consultants, advisors, designers and overseers of accounting processes. Borrego, Pardal, and Carreira (2020) conclude that accounting is one of the professions that is undergoing the most profound technological changes. Successful technology implementation is a prerequisite for delivering high-quality services and maintaining a competitive edge and market position, particularly during the pandemic. It also catalyzes change in how accounting processes are performed (Kaka, 2020). Özdoğan (2017) notes that we may see a positive impact of informational technology in performing everyday accounting tasks. Today, it is impossible to perform accounting duties without the support of information technologies. Technology also impacts the quantity and quality of financial information, resulting in a more effective financial reporting system and a more optimal capital allocation.
Furthermore, AI technologies can automate simple and repetitive accounting tasks, which allows accountants to allocate more time to complex issues (Rîndaşu, 2017). Consequently, accountants can potentially create more added value for the organization, which was previously impossible. This also means that AI software will eliminate or replace some accounting processes that humans perform today. However, currently, the complete replacement of human labor seems impossible due to the required expert knowledge in local law.
There are many modern technologies being implemented in accounting departments, such as Big Data, robotics, automation, blockchain, etc. However, AI technology seems to be unique. It integrates other technologies, further reducing the costs of accounting operations obtained as a result of using more elementary technologies, such as automation and robotics. Notably, AI in accounting not only makes the accounting processes more efficient but also introduces a necessary element of innovation in the processes being implemented (Ghani, Ariffin, & Sukmadilaga, 2022). Artificial intelligence supports a very important financial management tool that is critical for financial and accounting departments, such as forecasting, financial planning, pro forma financial statements and fraud detection. The quality of accounting data has also increased (Al Wael, Abdallah, Ghura, & Buallay, 2023). Moreover, AI has the potential to improve operational, strategic and other aspects of a company's operations that other technological advances may not be able to handle (Rawashdeh, Bakhit, & Abaalkhail, 2023). Artificial intelligence enables more effective control processes and generates crucial information necessary for making key decisions in the company (Anh et al., 2024). However, implementing AI in the work of an accountant is associated with numerous challenges. Accountants commonly consider the implementation process itself to be complex, complicated, expensive and requiring advanced IT systems (Ghani et al., 2022). Among other barriers, Al Wael et al. (2023) mention the lack of professional knowledge or advice on the subject, concerns about the AI's impact on the number of people employed, the lack of openness of accountants to changes, the lack of understanding of AI among managers and ethical and formal problems. The constantly changing accounting standards and tax laws are also significant.
The issue of AI implementation is interesting from the theoretical perspective, i.e. resource-based view (RBV) theory or diffusion theory (DT). As Barney (1991) advocates, the RBV theory assumes that resources that are valuable, rare, inimitable and nonsustainable (VRIN) are key to competitive advantage. The RBV theory refers to unique resources at the firm's disposal, such as human capital, organizational culture and know-how. It perceives them as the long-term drivers of the company's survival and success. Many studies focused on such areas as business information and networking (Sampler, 1998). Moreover, AI technologies are transforming the landscape and may render existing resources obsolete or create new ones, particularly when processing accounting information. The managerial perception is key in determining whether, in the context of the accounting process, AI technology is important.
The diffusion theory addresses the issue of social acceptance of innovation. Rogers (1995) lists factors influencing the rate of technology acceptance and diffusion: complexity (the degree of difficulty in using), compatibility (the degree of consistency and continuity), relative advantage, observability (the ability to discern benefits) and trialability (the ease of using the innovation). Therefore, our study contributes to the DT by examining how accounting practitioners adopt AI technology.
The technology-organization environment (TOE) framework also exists in the literature and assumes that the implementation of technology in a given organization depends on factors of a technical, organizational, and environmental nature. It enables researchers to categorize factors that influence the adoption of new technologies in organizations (Ghani et al., 2022).
The technology acceptance model (TAM) also warrants mention. According to the model's assumptions, behavioral intentions determine whether a given organization will accept the technology. It is the end-users of the technology who decide on its acceptance, and its most important criteria should be: usefulness in performing individual processes and perceived ease of use. Effective communication of new technology facilitates its implementation and acceptance (Al Wael et al., 2023).
Moreover challenge and response theory can facilitate the practical implementation of technology (Rawashdeh et al., 2023). When a challenge arises, organizations may be unprepared and not ready to use a given concept (technology), so they will strive to reject it. However, the opposite situation is possible, in which the company responds positively to challenges, adapting its current activities to new expectations. In the case of AI implementation, these will be additional financial resources allocated for AI implementation and necessary training.
In the literature, we may distinguish another approach, called technology readiness (TR) (Anh et al., 2024). The most important aspect that determines the use of a given technology is its maturity, so that it can be successfully developed, introduced and its practical application can be completed properly. Here, we should distinguish between the development phase, related to the prototype phase, testing and development, and the operational phase, which we may define as readiness for effective functioning and credibility in the operational activity of the enterprise.
As a profession, accountants constitute a very interesting population from a research perspective. According to the United Nations Conference on Trade and Development (UNCTAD) Secretary-General, today's accountants are in times of unprecedented and irreversible technological change from Accounting 4.0 to Accounting 5.0 (United Nations, 2018). The survival of the accounting profession depends on how the accounting community responds to the technological challenge (Borrego et al., 2020). Experts from McKinsey & Company predict that in the near future, the profile of an accounting professional will require a deep understanding of the digital environment [1]. This relates to the faster information flow within and outside the organization, as well as the ability to implement more innovative accounting processes, which leads to a more effective accounting system (Nimigean, 2021).
The study aimed to empirically investigate the factors influencing the use of AI in the accounting profession in Poland. The COVID-19 pandemic hastened modernization in almost all economic sectors. However, the accounting sector is exceptional because it reinforces existing and strong trends. Poland is a special case because it is one of the most developed markets in the outsourcing sector in the world. It stems from specific characteristics, such as relatively low labor costs, well-educated employees, and a favorable geographical location. The market for accounting services is developing dynamically. In particular, large city centers should have appropriate university facilities to educate future professional staff. According to the Association of Business Service Leaders in Poland (ABSL) report (ABSL, http), at the end of the first quarter of 2021, over 1,600 Polish and foreign centers in the business services sector in Poland provided accounting services, among others. Most outsourcing centers (56%) serve global clients, around 40% are dedicated to selected countries or regions, and only 2.6% operate vis-à-vis domestic enterprises. Approximately 60% of the processes handled by outsourcing centers involve accounting services. In Business Process Outsourcing (BPO) centers operating on a global scale, scholars have noticed the increasing importance of intelligent process automation (IPA). In Poland, the robotization process is just beginning, and the COVID-19 pandemic has significantly contributed to the automation of accounting processes.
The pandemic served not only as a catalyst for technological transformation, increasing the pace of changes, but also set it on a specific trajectory. In the case of the accounting profession, it led to working under a home office regime, imposing the necessity for self-discipline and self-reliance. Face-to-face communication has transitioned into the digital space. For accountants, the customs of work have been irreversibly reshaped. Although the pandemic has subsided and pre-pandemic conditions have largely resumed, some behavioral patterns established during that period have remained. Today, we witness a notable shift toward flexible work locations, with accountants being more integrated into virtual workspaces than physical ones.
Researchers conducted previous studies on factors influencing the use of AI in Malaysia and concerned public companies (Ghani et al., 2022), Kuwait in the public sector (Al Wael et al., 2023), the Middle East in the small and medium-sized enterprise sector (Rawashdeh et al., 2023), Vietnam (Anh et al., 2024) and listed companies in the United States (Iwuanyanwu, 2021). We contribute to the existing literature by examining factors that no one has analyzed to date, i.e. the impact of remote work and the changes in business models that affect the use of AI in accounting departments.
We divided the article into the following sections: new technologies in the accounting profession, research design and results. In the last section, we present concluding remarks.
New technologies in the accounting profession
The accounting profession is undergoing a technological revolution. The scale and speed of the change depend on many factors, i.e. country, sector and company type. The transformation seems inevitable and will ultimately impact the accounting profession. The study by Rîndaşu (2017) provides empirical evidence that corroborates the hypothesis positing that information technology has a profound impact on the accounting environment. The research involved questionnaires collected from 115 accountants and financial auditors working in Romania. Most respondents exhibited an average level of theoretical knowledge of information technology. The results suggest that the implementation of AI technologies to support accounting and auditing processes has already begun. At the same time, respondents were aware of the risks and threats related to applying AI technologies in accounting. According to the expressed opinions, an attitude based on lifelong learning (i.e. course participation) is necessary to be up-to-date with new technologies. It also requires constant development of university curricula and updating of content and training programs conducted by professional organizations of accountants and financial auditors.
Based on their research, Zenuni, Begolli, and Ujkani (2014) conclude that knowledge and skills related to new technologies are crucial. Their questionnaire results indicate that most accountants already have basic knowledge of IT environments, which is necessary for day-to-day accounting work. However, the study also reveals that after graduation from university, accountants rarely have the opportunity to educate themselves on new technologies. More sophisticated and detailed questions uncovered that accountants were unaware of the potential and capabilities of modern AI technologies in accounting. Zenuni et al. (2014) recommend that accountants allocate more time to training related to IT technologies and the changes in university curricula, with an emphasis on AI technologies and their applications in accounting.
The accounting profession is facing the challenge of implementing not one, but a whole group of technologies, labeled artificial intelligence (Yap & Drye, 2018; Zawacki-Richter, Marín, Bond, & Gouverneur, 2019; Hwang, Xie, Wah, & Gašević, 2020). Duan, Edwards, and Dwivedi (2019) provide the following examples of areas where AI technologies can be useful: decision-making, financial analysis, estimation, judgment, prediction, natural language processing and recommendation. Other application areas related to AI are security, technology democratization, hyper-automation and deep learning techniques, such as image recognition and voice recognition. Some AI technologies have already entered the accounting profession or are expected to have a significant impact in the near future, including machine learning, robotics, data mining, big data and expert systems. We may see these technologies as branches growing from the same root. We need a closer look at new technologies to understand the challenge of innovation.
Automation will first replace routine and repetitive tasks. Robotics is one of the key AI technologies, defined as the software that enables robots to perform repetitive tasks (Ashok, N, & M.S, 2019). In accounting, it is visible in recording sales and financial transactions, sorting documents and preparing financial statements (Seasongood, 2016). Machine learning (ML) is already present in accounting, especially in large international corporations. Das, Dey, Pal, and Roy (2015) define ML as the computer's ability to learn without specifically dedicated software. It is based on algorithms that perform data analysis and make decisions, followed by an appropriate course of action. Today's ML application relates to formulating proper accounting treatment, predicting market prices, searching and indexing documents, and analyzing databases. Moreover, organizations employ ML techniques in bankruptcy prediction (Barboza, Kimura, & Altman, 2017; Qu, Quan, Lei, & Shi, 2019) and accounting estimates (Cho, Vasarhelyi, Sun, & Zhang, 2020).
Another key technology important for accounting professionals is Big Data. It is not easy to define because one understands it differently depending on the company's size. The perception of the database size depends on the capabilities of the information system in terms of storage and processing abilities. Zhang, Xiong, Xie, Fan, and Gu (2020) identify important Big Data qualities as the five Vs. We may decode it as a massive Volume of the database, a high-Velocity of data continuously, a large variety of data types and uncertain Veracity, which addresses the issue of data accuracy and reliability, and Value, which examines cost-benefit criteria of the data collection process.
Large databases offer the potential to uncover statistical correlations between employees' actions and behaviors and company effectiveness. For example, the tone of a telephone conversation or e-mail may indicate the staff's morale and motivation. Another example of Big Data use is the so-called “beyond budgeting,” which considers the usage of sources other than traditional information, such as satellites, climate, environment, macroeconomy data, etc. In the case of financial accounting, Big Data allows the enrichment of accounting records with multimedia information. It is especially useful for assets like property, plant & equipment or investment properties, where pictures and movies may be more informative and persuasive than numbers. Consequently, accounting information is more transparent, reliable and credible for users. It provides more detailed data for accounting estimates, such as fair value measurement (especially when market prices are unavailable). For the same reason, it may be useful for financial auditors (Warren, Moffitt, & Byrnes, 2015).
The impact of AI technologies like Big Data, ML and others on business is twofold. Firstly, it changes the business model, and accounting is trying to follow up on changes and adopt the accounting policy accordingly. It results in finding adequate measurement methods for new types of transactions. Secondly, a more direct impact is visible in the performance of accounting tasks, such as the automation of transaction settlement (receivables and payables), automated accounting records and a more automated process of financial audit, among others. Krahel and Vasarhelyi (2014) note that the accountant's role evolves from the person responsible for recording transactions and preparing financial statements to the expert monitoring and explaining the accounting process. The accountant's attention is shifting toward understanding and processing financial data and taking care of the reliability of the financial statement. The emergence of AI technologies brings new stakeholders to the table, namely designers and servicemen of accounting systems (Al-Htaybat, Alberti-Alhtaybat, & Alhatabat, 2018). The implementation of AI in accounting departments is unique for at least two different reasons and may change the work of accountants to a great extent. First, AI is the most technologically advanced solution and requires other, easier technologies to be introduced in advance, such as automation and robotization of accounting processes. Second, it integrates and allows for the management of other technologies, such as blockchain, Big Data, creating one coherent IT environment for accountants.
Research design
Research process
From both the single-entity perspective and the broader economic perspective, the key issue is identifying the drivers of technological change. The crisis caused by COVID-19 created numerous health, sociological and economic threats. On the other hand, the pandemic accelerated changes in many areas of human behavior. The main goal of our study was to identify the key drivers of AI implementation in the accounting profession.
The study also allowed us to evaluate to what degree digital transformation changed the business model and accounting practice of companies, corporations, and public institutions in Poland. We based the empirical analysis on the results of the extensive survey conducted in the year 2021 among 231 corporate managers of small, medium, large and multinational corporations. The population varied in many aspects, i.e. in small and medium entities, the accounting differed in terms of the scope, regulations, problems and organization as compared to the large and multinational companies. However, we assumed that the core accounting activities and their nature remained homogenous, which was processing financial information. Simultaneously, we may consider accounting to be the backbone of each organization and essential for each manager, with no regard for department affiliation. We used the Paper & Pen Personal Interview method, and collected the survey during MBA classes.
We consider COVID-19 to be a context variable. We assumed that digital transformation and technological progress following the pandemic would create momentum to change the business models of companies, corporations and public institutions in Poland, and consequently, the accounting practices. We defined technological progress as implementing AI technologies like digitalization, automation, robotics, Big Data, machine learning and others, where the Internet plays a crucial part. These technologies mainly collect, process, manage and exploit data.
Research design and hypothesis development
There are a few empirical studies in the literature that analyze the factors influencing the use of AI in accounting departments. Ghani et al. (2022) divided the most important factors influencing the implementation of AI into the following groups: information technology (IT), top management support and government support. Al Wael et al. (2023) distinguished factors on the organizational level: organizational culture, communication structure, competitive pressure, regulatory support, resources and individual level: perceived usefulness, perceived ease of use. Iwuanyanwu (2021) proposed the following factors influencing the implementation of AI: financial institutions/fund providers, a company's customers, multinational organizations, government regulations, a company's shareholders and recommendations of consultants/professional bodies. In turn, Yang, Blount, and Amrollahi (2021) analyzed the process of implementing AI in large accounting firms. The authors considered the following factors as the most important in facilitating the implementation: communication process, regulatory environment, predicted industrial changes, client's acceptance, firm scope, technology affordance and technology barriers.
Rawashdeh et al. (2023) adopted the perspective of the last AI user and considered the most important determinants to be: technological compatibility, readiness for challenge, saving time and efficiency-improving, while Anh et al. (2024) empirically verified mainly technology readiness, and the perceived ease.
The main variable of interest is management's attitude (opinion) about the most necessary changes to be implemented in the entity's activity. More specifically, we coded the dependent variable as one in the case of companies in which managers opted for AI as a priority to be implemented in accounting and zero otherwise. We selected AI because it encompasses the whole set of the newest technologies. We conjecture that managerial awareness concerning AI's potential and advantages is critical in their implementation and adaptation decisions. We must stress that in the survey, managers could have opted for other areas of AI implementation than accounting. Therefore, the choice manifests in the managers' reasoning based on their experience and expectations. We suspect that it will also influence the course of action managers undertake in the future.
Our findings identified companies where AI played a crucial role in accounting or at least was expected to do so soon. Our study also considered a time factor. For this reason, we included COVID-19 by introducing issues related to the situation before and during the pandemic in the questionnaire. Findings from other authors based on the same data source identified key characteristics influencing managerial decisions, such as the one that is our key variable of interest. For this reason, we used a set of independent variables whose impact on the managerial decision was already documented: market coverage (regional, local, nationwide, international), expected change of business model after the pandemic (a dichotomous variable, 1 – substantial change, 0 – no change), the percentage of employees who used a remote work before the pandemic, the percentage of employees within the organization who were working only in a “distance mode” during the pandemic, the degree to which employees' competencies met managers' requirements, the impact of the employees' education level on the ease of “remote work” adaptation during the pandemic.
Moreover, we also introduced two moderating variables, i.e. the organization's size and sector affiliation. The number of employees proxies the organization size, and it seems to be the most exact size measure in this case. Other studies adopted a similar approach (Bakarich & O'Brien, 2021). Our sector profile affiliation variable classified companies into three groups: production, services and other. We suspected that these variables would moderate the impact of independent variables. In other words, we conjectured that the COVID-19 impact on accounting varies depending on the organization's size and sector profile.
The significance of remote work increased during the pandemic. Consequently, accounting activities involve more and more information technologies (Saxena & Khandelwal, 2022). According to Y. E. Fukumura, Gray, Lucas, Becerik-Gerber, and Roll (2021), remote work requires the application of a wider spectrum of dedicated software and working time with the computer increases on average by 1.5 hours a day. Similarly, Joamets and Chochia (2020) note that from the beginning of COVID-19, people began performing many accounting tasks in remote mode. In response, it created additional demand for further digitalizing of accounting processes. Therefore, we suspected that the greater the implementation of remote work processes in accounting departments, the greater the need to introduce modern technological solutions, which may facilitate the introduction of more advanced AI technologies. Remote work is becoming an everyday reality in accounting and other professions. Furthermore, AI technologies have the potential to support remote work, make it more effective and provide control and audit tools. Consequently, continuous investment in AI technologies seems inevitable (Leonardi, 2021; Patil & Gopalakrishnan, 2020). However, others claim that there will be a massive return to stationary work after the pandemic, reducing demand for modern technologies.
Moreover, AI technologies pressure accountants to acquire new competencies and skills (Stancheva-Todorova, 2018). Close cooperation between accountants and computer specialists is essential to implement AI technologies effectively. It is the only way to solve more advanced accounting problems. This type of cooperation emerged in many companies during the pandemic. According to Jaiswal, Arun, and Varma (2022), the prerequisites for AI implementation are adequate experience and cognitive and technological skills. Remote work pushed accountants to acquire these skills and competencies. Moreover, during the pandemic, many accountants learned and recognized the usability of AI technologies or their indispensability in everyday work, which Wassan (2021) also noted.
To summarize, voluntary remote work before the pandemic translates to greater acceptance of advanced technology, more employee experience and skills, and better organizational preparation for AI implementation. On the other hand, organizations that did not invest in new technologies before the pandemic are currently not prepared for AI implementation. Considering the premises of the previously addressed technology-organization-environment (TOE) framework and the technology acceptance model (TAM), we anticipated that readiness and openness to technological change were crucial drivers of successful AI implementation. Therefore, we conjectured the following:
Managers of organizations (firms, NGOs, public companies) that, before the pandemic, were using remote work (REM) are more ready and willing to implement AI technologies in accounting after the pandemic. The size (SIZE) and profile (PROF) of the organization moderate the impact.
To a greater extent, entities (firms, NGOs and public companies) started using remote work methods during the pandemic (DIST), the fewer managers of those organizations are willing to implement AI technologies in accounting after the pandemic. The size (SIZE) and profile (PROF) of the organization moderate the impact.
According to Soni et al. (2019), there was a natural tendency for AI implementation in technologically advanced organizations, whose business model is strongly tied to the application of new technological advancements. Innovation and global competition put pressure on the business model to change and, consequently, induce the implementation of AI technologies. Therefore, fast and successful AI implementation may lead to a strategic technological advantage. Norman (2017) presents a similar view on the issue.
Another aspect is the change that emerged in the market during the pandemic. The clients are driving these changes, and their expectations pertain to products and services. This may require a change of business model based on AI implementation, especially concerning advanced production process automation. In some cases, numerous modifications in process production may necessitate a change in business model, which is increasingly tied to technology investments, especially in AI.
A company's success continually depends on effectively realizing innovative economic processes and selling innovative products and services (Di Vaio, Boccia, Landriani, & Palladino, 2020). The prerequisite for achieving the goals mentioned above is creating a business model based on innovation. Furthermore, the introduction of AI has the potential to further enhance the enterprise's efficiency and competitiveness. According to Strusani and Houngbonon (2019), AI enables a positive market shift for enterprises, offering a wider assortment and more intense interaction with the market. The introduction of an innovative business model provides the potential to increase sales and, at the same time, reduce costs and market entry barriers.
Brooks, Gherhes, and Vorley (2020) identify other external factors that necessitate a change in the business model and, consequently, AI implementation. The external pressure stems from changes in the industry and the closest company environment. Companies already using AI technologies also create market pressure to change business models. Another noteworthy factor is the internal technological barriers present in each enterprise. The findings of Brooks et al. (2020) suggest that even though economic enterprises are aware of internal and industry barriers, they are very difficult to overcome. Notably, the change from the traditional business model is not easy, as there is strong resistance to the implementation of new business models. Many accounting companies underrate and neglect innovation factors in the accounting profession, even though advanced technologies may provide many benefits and significantly increase the effectiveness of accounting processes.
Last but not least is the organizational culture. Internal factors related to organizational issues may significantly influence the creation of innovative business models based on AI technologies (Lee, Suh, Roy, & Baucus, 2019). In this regard, important factors are organizational culture, organizational values and organizational design. The creation of an innovative culture is the opposite of a more traditional or even archaic business model, which is often based mainly on leadership. All in all, accountants and CFOs are primarily responsible for implementing AI in economic entities.
We may identify the opposite relationship, in which AI technologies are the main actors and can improve market position, ease innovation processes and intensify R&D activity, etc., which, as a result, forces a change in business model (Soni et al., 2019). According to the findings of Toniolo, Masiero, Massaro, and Bagnoli (2020), the implementation of AI plays a crucial role in creating sustainable business models.
Based on the literature review, with a special reference to the Challenge and Response Theory, we should note that the change in the business model puts pressure on technological change, increasing the employees' competencies, acceptance of changes by the workforce and change of the organizational culture into predevelopment and pro-technological, which creates an avenue for AI adoption. Based on that, we conjectured the following hypothesis:
Organizations planning to change their business models after the pandemic (B_MOD) are less willing to adopt AI technologies. The size (SIZE) and profile (PROF) of the organization moderate the impact.
We employ models 1 and 2 to test H1, models 3 and 4 to test H2, and models 5 and 6 to test H3.
A summary of variables is presented in Table 1, and descriptive statistics for the analyzed variables in Table 2.
Summary presentation of variables
| Variables | Abbreviation | Description |
|---|---|---|
| Dependent variable | ||
| AI implementation | AI | managers of organizations opting for the implementation of AI technologies in the near future in accounting (a dichotomous variable) |
| Independent variables | ||
| Remote work | REM | the percentage of employees who used remote work before the pandemic |
| Distance mode | DIST | the percentage of employees within the organization who are working only in a “distance mode” during the pandemic |
| Business model | B_MOD | expected change of the business model as a result of the pandemic |
| Control variables | ||
| International market coverage | IMC | international market coverage (a dichotomous variable) |
| Nationwide market coverage | NMC | nationwide market coverage (a dichotomous variable) |
| Public companies | PC | public companies (a dichotomous variable) |
| Municipal companies | MC | municipal companies (a dichotomous variable) |
| Other companies | OC | other companies (a dichotomous variable) |
| Employees' competencies | COMP | the degree to which employees' competencies meet managers' requirements |
| Basic education | B_EDU | the impact of the employees' education level on the ease of “remote work” adaptation during the pandemic – basic and vocational education (a dichotomous variable) |
| Secondary education | S_EDU | the impact of the employees' education level on the ease of “remote work” adaptation during the pandemic – secondary education (a dichotomous variable) |
| Higher education | H_EDU | the impact of the employees' education level on the ease of “remote work” adaptation during the pandemic – higher education (a dichotomous variable) |
| Moderating variables | ||
| Size | SIZE | the organization’s size as proxied by the number of employees |
| Profile | PROF | the organization’s profile (the variable may take three possible values: production, services and other) |
Descriptive statistics
| Variable | N | Mean | SD | Median | Min | Max | Q1 | Q3 |
|---|---|---|---|---|---|---|---|---|
| Dependent variable | ||||||||
| AI | 230 | 0.38 | 0.49 | 0.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| Independent variables | ||||||||
| REM | 230 | 0.62 | 0.49 | 1.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| DIST | 230 | 2.32 | 1.35 | 2.00 | 1.00 | 5.00 | 1.00 | 3.00 |
| B_MOD | 230 | 0.50 | 0.23 | 0.55 | 0.10 | 0.95 | 0.55 | 0.55 |
| IMC | 231 | 0.29 | 0.45 | 0.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| NMC | 231 | 0.44 | 0.50 | 0.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| PC | 231 | 0.18 | 0.38 | 0.00 | 0.00 | 1.00 | 0.00 | 0.00 |
| MC | 231 | 0.08 | 0.28 | 0.00 | 0.00 | 1.00 | 0.00 | 0.00 |
| OC | 231 | 0.09 | 0.29 | 0.00 | 0.00 | 1.00 | 0.00 | 0.00 |
| COMP | 230 | 0.33 | 0.47 | 0.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| B_EDU | 230 | 0.50 | 0.27 | 0.33 | 0.06 | 0.95 | 0.33 | 0.72 |
| S_EDU | 230 | 0.50 | 0.26 | 0.30 | 0.02 | 0.95 | 0.30 | 0.73 |
| H_EDU | 230 | 0.50 | 0.27 | 0.50 | 0.07 | 0.95 | 0.31 | 0.70 |
| Moderating variables | ||||||||
| SIZE | 231 | 0.70 | 0.46 | 1.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| PROF.−1 | 230 | 0.31 | 0.46 | 0.00 | 0.00 | 1.00 | 0.00 | 1.00 |
| PROF.−2 | 230 | 0.22 | 0.42 | 0.00 | 0.00 | 1.00 | 0.00 | 0.00 |
Results
The logistic regression results (Table 3) suggest that the main variables of interest, REM, DIST, and B_MOD, significantly influenced the dependent variable. First, the more organizations used “remote work” before the pandemic, the more willing they were to adopt AI technologies in accounting. We suspect that it concerns organizations characterized by a highly computerized environment. In this case, the pandemic situation catalyzed the decision to adopt AI technologies. Therefore, the results in this regard are not surprising. The literature also emphasizes the technical possibilities of introducing AI. Moreover, the more stakeholders in the enterprise are involved in the AI implementation process, the greater the chance of success (Ghani et al., 2022). Anh et al. (2024) emphasize the need to achieve technological readiness in the enterprise, without which the implementation of the discussed solutions will not be possible. The obtained results are consistent with the resource-based theory and the technology-organization environment (TOE), enterprises with specific IT and organizational resources can successfully implement AI.
Results of the logistic regression with the odds ratio
| Variables | OR | 95%CI | p | |
|---|---|---|---|---|
| REM | 1.235 | 1.140 | 1.338 | <0.001 * |
| DIST | 0.906 | 0.871 | 0.942 | <0.001 * |
| B_MOD | 0.730 | 0.607 | 0.878 | 0.001 * |
| IMC | 1.009 | 0.825 | 1.233 | 0.934 |
| NMC | 1.054 | 0.876 | 1.267 | 0.579 |
| PC | 1.297 | 1.054 | 1.596 | 0.015 * |
| MC | 0.493 | 0.376 | 0.647 | <0.001 * |
| OC | 1.145 | 0.993 | 1.320 | 0.064 |
| COMP | 0.899 | 0.823 | 0.982 | 0.018 * |
| B_EDU | 3.341 | 2.580 | 4.327 | <0.001 * |
| S_EDU | 1.445 | 1.102 | 1.893 | 0.008 * |
| H_EDU | 0.691 | 0.558 | 0.856 | 0.001 * |
Note(s): *significance at 5% level
Second, the more employees worked remotely during the pandemic, the less they were willing to adopt AI technologies. We suspect that one can explain this by the fact that COVID-19 forced these organizations to make rapid changes toward remote work. Simultaneously, these organizations failed to establish an adequate organizational and pro-technological culture. Therefore, they were not ready to undertake further challenges in this regard. Therefore, without additional financial and organizational investments, implementing AI in these enterprises will be challenging, which is consistent with the challenge and response theory (Rawashdeh et al., 2023). Al Wael et al. (2023) obtained similar results. In their opinion, the right organizational culture supports the implementation of complex technological solutions, such as AI. However, it is not the only factor determining the successful implementation of AI. An equally important element is the perception of technology as easy and useful, which in turn confirms the diffusion theory and the technology acceptance model. Effective communication between management and employees is crucial in this regard (Al Wael et al., 2023). In a similar vein, Rawashdeh et al. (2023) also provide empirical evidence, suggesting that the impact of organizational culture on the AI implementation process is equally important, as the convergence of current procedures and practices with AI is emphasized.
The findings also demonstrate that organizations, which, to a smaller extent, plan to change their business model after the pandemic, are more willing to adopt AI technologies in the near future. Therefore, we suspect that organizations that are more adaptive and flexible and are better equipped to withstand the economic consequences of COVID-19 are those that are more prepared for the adoption of AI technologies. Rawashdeh et al. (2023) obtained similar results. The authors indicate that not all organizations fully accept changes. Hence, the implementation of technological solutions may also be problematic for them. Table 3 presents the results of the logistic regression with the odds ratio.
The implementation of the moderating variables would shed more light on this issue. To capture the effect of the organization's size and profile, we introduced moderating variables SIZE and PROF. In the following (Tables 4–9), we present the moderating effects. We observed no moderating effect of SIZE on REM (see Table 4).
Moderating effects of SIZE on REM
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| REM | 1.205 | 1.076 | 1.349 | p = 0.001 * | 1.205 |
| SIZE*REM | 1.037 | 0.925 | 1.161 | p = 0.535 | 1.250 |
Moderating effects of PROF on REM
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| REM | 1.408 | 1.223 | 1.621 | p = 0.001 * | 1.408 |
| PROF.−1*REM | 0.851 | 0.736 | 0.983 | p = 0.029 * | 1.198 |
| PROF.−2*REM | 0.835 | 0.683 | 1.02 | p = 0.079 | 1.176 |
Moderating effects of SIZE on DIST
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| DIST | 0.932 | 0.888 | 0.977 | p = 0.004 * | 0.932 |
| DIST*REM | 0.966 | 0.934 | 0.999 | p = 0.046 * | 0.900 |
Moderating effects of PROF on DIST
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| DIST | 0.927 | 0.889 | 0.967 | p = 0.001 * | 0.927 |
| PROF.−1*DIST | 0.929 | 0.881 | 0.979 | p = 0.007 * | 0.861 |
| PROF.−2*DIST | 0.937 | 0.881 | 0.997 | p = 0.04 * | 0.869 |
Moderating effects of SIZE on B_MOD
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| B_MOD | 0.772 | 0.639 | 0.933 | p < 0.008 * | 0.772 |
| SIZE*B_MOD | 0.837 | 0.715 | 0.981 | p = 0.029 * | 0.646 |
Moderating effects of PROF on B_MOD
| Variable | AOR | 95%CI | p | Total AOR | |
|---|---|---|---|---|---|
| B_MOD | 0.674 | 0.466 | 0.976 | p = 0.038 * | 0.674 |
| PROF.−1*B_MOD | 1.009 | 0.724 | 1.405 | p = 0.958 | 0.680 |
| PROF.−2*B_MOD | 1.285 | 0.875 | 1.887 | p = 0.203 | 0.866 |
We detected an interaction between PROF and REM (see Table 5). In this case, we used the moderating variable in three possible states: production (PROF−1 = 1 and PROF−2 = 0), services (PROF−1 = 0 and PROF−2 = 1) and others (PROF−1 = 0 and PROF−2 = 0). The results suggest that the impact of REM was the strongest in production and weaker but still significant in services.
The variable SIZE moderates the impact of DIST on the dependent variable (Table 6). The effect is the strongest in large organizations and weaker in smaller ones. We suspect that large organizations in which, to a larger extent, employees were working only remotely during the pandemic have already implemented AI technologies in accounting, which provides additional empirical argument supporting our previous conclusion.
We also observed the moderation effect of PROF on DIST (Table 7). The results suggest that the impact of DIST on the dependent variable was the strongest in services and weaker but still significant in production. We conjectured that organizations in the service industry were more apt to remote work (i.e. contact with customers, providing services online, etc.). For this reason, there is a greater probability that AI technologies had been implemented in these companies before the pandemic. The effect on organizations in the production industry was much weaker. It suggests that these organizations were still in the dawn of the AI era.
The next variable of interest is B_MOD, which captures the management's expectation of the business model change after the end of the pandemic. The results suggest (Table 8) that managers of organizations who expect the business model to change were less willing to adopt AI technologies in accounting. We conjectured that there were more pressing needs for organizations most severely affected by the COVID-19 consequences. The impact of B_MOD was negative and significant for small, medium and big organizations. However, more prominent for the latter group. The results suggest that a crisis may cause trouble and delay the organization's modernization process.
We found no interaction between PROF and B_MOD, suggesting that the type of organization (production, service and others) does not moderate the negative impact of B_MOD on the dependent variable (Table 9).
Moreover, we also used a set of control variables. Our findings suggest that public (municipal) organizations were more (less) willing to adopt AI technologies. In Poland, people perceive municipal entities technologies as more traditional and inertial, with a larger degree of resistance to change. Therefore, it is unsurprising that these organizations resist novelties and stick to the past. We observed a competitive approach in the case of public organizations, like public companies. Managers seek new ways of doing business and are looking for cost savings. Thus, the results are logical and expected.
We also controlled for the impact of employees' education levels on the dependent variable. The results are interesting and suggest that managers of organizations with employees representing basic, vocational and secondary education levels are more willing to adopt AI technologies in accounting. We could explain it by the fact that AI technologies are more apt to replace simple and repetitive tasks. For this reason, implementing AI technologies is much easier and more economical. In the case of a highly educated workforce (tertiary education), it is referred to as the expert's more sophisticated, unique work, which is much more difficult for AI to imitate. Our findings align with the conclusions of the empirical study by Ghani et al. (2022) regarding the impact of employee education on the AI implementation process.
We tested the logistic regression models for multicollinearity by checking Pearson coefficients between variables and VIF. The results did not imply collinearity.
Concluding remarks
The accounting profession has undergone significant changes recently. Accountancy is believed to be one of the professions expected to face the most fundamental challenges. Increasingly, information technology determines the efficiency of accountants' work and accounting departments. They improve the speed of information processing and make financial information more reliable and useful for the end user. The application of AI technologies results in a reduction of errors in an accountant's work and fewer distortions in financial statements. An accountant is transitioning from a role focused on recording transactions to one that involves making strategic decisions or preparing the financial basis for informed decisions, which may enhance the company's competitive advantage. The scope of qualities required in the accounting profession is changing. In the past, a successful accountant should have known detailed regulations and laws and should have been reliable and honest. Today, new competencies and skills are invaluable, like the adaptability to new technologies, the ability to work in a digitalized environment, designing e-models and providing consulting and advisory services based on large databases.
During the COVID-19 pandemic, in many companies, we could observe the acceleration in the adoption of remote work in the accounting departments. The best-prepared companies for technological change were those that had already used a remote work mode before the pandemic. Other companies were forced to enter remote work mode due to governmental restrictions. During the pandemic times, numerous problems emerged in the accounting work, such as difficult access to accounting documentation, troublesome communication within the team, dependence on the work performed by other team members, difficulties in making decisions, etc. The best solution for the above-mentioned problems was to employ the achievements of modern technologies. However, access to modern technologies is not always possible for every company.
In many companies, the digitalization of accounting processes is already advanced. However, this is not the case in other areas of the company's operations. The usability potential seems to be the greatest in accounting. It applies to technologies like Big Data, Machine Learning, robotics and Data Mining, which we classify as AI technologies. It allows for more effective and faster decision-making processes. In more detail, it allows for deeper analyses, more precise accounting estimates (based on large databases), faster accounting processes and error detection. Consequently, it leads to a more effective accounting system generating financial information of a higher quality. It is particularly useful in repetitive and standardization-prone tasks. Finally, AI offers further opportunities to improve the security of accounting data due to technologies like hyper-automation, deep learning techniques and voice recognition.
Our study provides an empirical argument supporting H1. Therefore, we conclude that managers of organizations (firms, NGOs, public companies) that, before the pandemic, were using remote work are more ready and willing to implement AI technologies in accounting after the pandemic. The impact is moderated only by the type of organization, but not by the size. We suspected that managers already acquainted with remote work would also be more aware of the AI advantages, and these organizations, before the pandemic, would already be better prepared for the modernization in accounting. The findings corroborated our expectations, according to which companies ready to use remote work mode in the area of accounting processes are those that are more technologically advanced. In those companies, accountants before the COVID-19 pandemic acquired technological skills and competencies and were already aware of the benefits of more advanced technologies. At the same time, they were willing to accept more complex AI technologies. We believe that creating an adequate organizational culture and attitude among employees, supported by a well-designed training system, will lead to a more successful and faster AI implementation.
Our findings suggest that managers of organizations that have started using remote work methods during the pandemic were less willing to implement AI technologies in the accounting area after the pandemic. The impact is moderated by the size and profile of the organization (H2). The opinions of the surveyed also supported our H2. Organizations that had not used remote work mode before the pandemic usually do not have adequate tangible, intangible, technological and human resources. They were forced to introduce a remote work mode, which was not preceded by previous training in the use of technology in accounting in the pre-pandemic period. We suspect that in those companies, the digitalization of accounting processes was at a very low level.
Finally, our results suggest that the most affected organizations were those that planned the business model to change after the pandemic, and were the least interested in AI implementation in accounting. Therefore, our findings do not support H3, which means that the adoption of a modern business model based on a high saturation of activities with advanced technologies is not a sufficient condition for AI implementation. Similarly, organizational culture, organizational values and organizational design are not significant factors, which is contrary to what we expected. The possible explanation is that AI technology is a driver of business model change. Noteworthy, AI implementation allows for better analysis of the customer needs, better choice of operational or investment strategy, better management of the client's portfolio, and, as a result, the change of the business model.
To the best of our knowledge, the study is the first of its kind in CEE countries or Europe on the use of AI in accounting, particularly before and during the COVID-19 pandemic. This study contributes to the literature by providing a deeper understanding of AI adoption at the firm level, thereby filling existing gaps in the literature. It may strengthen the theories that underpin our understanding of the technology adoption by firms, revising, extending and elaborating the TOE framework with more empirical evidence. Our study demonstrates the conditioning of advanced technologies' implementation using the example of AI adaptation in the accounting profession. Furthermore, we analyzed new types of relationships before and during the COVID-19 pandemic, focusing on remote work practices, business model changes and the potential for AI implementation. To the best of our knowledge, this is the first study on this type of relationship. Moreover, unlike other studies, we considered factors dependent on the company and related to accountants at the same time. We provide empirical evidence suggesting that the organization's culture and the entity's openness to change strongly influence the use of AI in accounting processes, apart from technical issues. Finally, in the empirical part of the work, in addition to explanatory variables, we used moderating variables: the organization's size and organizational profile, which similar studies did not use.
Our findings suggest that the pandemic's impact on the organization's AI implementation was twofold. On the one hand, it forced faster modernization changes like implementing “remote work” techniques, a more computerized approach based on electronic documentation, customer service via the Internet, etc. It allows organizations to be better prepared, aware, and encouraged for AI implementation. On the other hand, for many organizations, pandemics caused economic problems, which delayed modernization investments in various areas of activity. Many entities did not perceive AI implementation in accounting to be the most urgent issue. Overall, we expect that AI implementation in other activities will facilitate similar changes in accounting. The divergent impact of the pandemic on AI implementation was also visible at the level of moderating factors. The company size and profile moderated the impact of independent factors of our model. Our results are consistent with the RBV theory and the diffusion theory, pointing to the need for unique resources, such as human capital or an appropriate organizational culture that promotes openness to changes in the AI implementation process, and indicating how AI technology spreads among accounting practitioners. Notably, TAM CRT may also have practical applications in the AI implementation process.
The primary challenge for academia in the future is to assess the transformative potential of AI. Our findings suggest that managerial readiness and openness to learn new technologies are key drivers of the AI implementation in accounting. We believe that it will reshape the landscape of the accounting profession. The accountants and companies that are the quickest to embrace technological change will dominate the market of the future. Organizations and professionals unwilling to change will eventually be pushed out. Artificial intelligence transforms the accounting profession in the way accounting tasks are performed, the way the accountant acquires information, and later on, acts based on that and solves the problems. Ultimately, it reshapes the role of the accountants, expanding the scope of their expertise. The process is currently underway, and we are observing its gradual progression.
Our findings will be useful for accounting and financial auditors in professional organizations. The use of even relatively simple technological solutions (remote work) accelerates the implementation of AI in accounting departments. If the company utilizes more advanced technologies and processes, such as robotics, implementing AI will be easier. Furthermore, the introduction of AI will enhance the efficiency of the technologies already implemented. If external factors, such as the COVID-19 pandemic, somehow force the current technological infrastructure to change, its impact on the implementation of AI will be smaller. An organizational culture that accepts the open-minded attitudes of accountants toward changes and technological advancements can also facilitate the use of AI. We should assess the technological competencies of employees similarly. However, if the company is undergoing dynamic changes in many dimensions, caused by, for example, the COVID-19 pandemic, a large scope of changes may mean that this is not the best time to implement AI in accounting departments. Accountants' experience in remote work, robotization, and automation of accounting processes facilitates the implementation of more advanced AI-type tools.
The research results may be helpful for managers who intend to implement more advanced IT solutions. Proper communication between the highest and lower levels of management also seems necessary in the discussed process. Making accountants aware of the significant benefits of implementing AI, such as time savings, operational efficiency and reduced accounting errors, will help reduce their reluctance to adopt AI. Companies should also provide more courses and training on modern technologies used in financial and accounting departments. It is also crucial to foster an organizational culture that values innovation, openness to change and new ideas. However, management must be aware that the initial outlays related to implementing AI may be high, while operational and cost efficiency, as well as a higher rate of return, will be observable in the long run.
Governmental institutions may also find our results interesting from a policymaking perspective. State policy should simplify the AI implementation process by providing subsidized training, tax benefits or expert knowledge. Similar solutions operate in the United States, which emphasize increasing employee competencies in addition to financial benefits. Consideration should also be given to providing special grants for companies interested in implementing AI in their financial and accounting departments. The success of such solutions in one company will drive the implementation of AI in other competitors in the industry.
Our study can also be useful for other scholars, because it sheds new light on the relationship between remote work modes before and during the COVID-19 pandemic in the accounting area, which scholars have not investigated previously. For educators, the research results indicate the need to change the curriculum. New subjects are necessary to cover issues related not only to financial and accounting computerized systems, but also to more advanced applications like “cloud computing,” Big Data, Machine Learning, Robotics, Data Mining and other AI technologies. The question is not whether the change will take place but when. The adoption of AI in accounting is inevitable, and the pandemic has only accelerated the pace. Only then will future graduates be fully ready to perform the function of an accountant successfully. It appears that due to the complexity and difficulty of the discussed phenomena, cooperation is required among the following groups: accountants, auditors, company managers, government institutions and schools.
Our study has at least several limitations. First of all, it is limited to Polish entities. The sample size was relatively small. We also did not control for several important characteristics, like the firm's profitability or leverage, the level of IT development, etc. This is because our sample consisted of various entities: public and private companies, NGOs and municipal companies. Therefore, future studies must focus on more coherent samples, which will enable control for more specific factors.
The authors would like to thank the anonymous reviewers and the editor for their insightful comments and constructive suggestions, which have greatly improved the article.
